Vous êtes sur la page 1sur 417

Qualitative Research in Action

Qualitative Research in Action

edited by

SAGE Publications
London • Thousand Oaks • New Delhi

Introduction  Tim May 2002
Chapter 1  Dorothy E. Smith 2002
Chapter 2  Sam Porter 2002
Chapter 3  Peter K. Manning 2002
Chapter 4  Christian Heath and Jon Hindmarsh 2002
Chapter 5  Malcolm Williams 2002
Chapter 6  Martín Sánchez-Jankowski 2002
Chapter 7  Nigel G. Fielding 2002
Chapter 8  Valerie Walkerdine, Helen Lucey and June Melody 2002
Chapter 9  Kathleen Gerson and Ruth Horowitz 2002
Chapter 10  Jennifer Mason 2002
Chapter 11  Steph Lawler 2002
Chapter 12  Linda McKie 2002
Chapter 13  Lynne Haney 2002
Chapter 14  Tracey Reynolds 2002
Chapter 15  Amanda Coffey 2002
Chapter 16  Lisa Adkins 2002
Chapter 17  Beverley Skeggs 2002
Chapter 18  Sherryl Kleinman 2002

First published 2002

Apart from any fair dealing for the purposes of research or private
study, or criticism or review, as permitted under the Copyright,
Designs and Patents Act, 1988, this publication may be reproduced,
stored or transmitted in any form, or by any means, only with
the prior permission in writing of the publishers, or in the case of
reprographic reproduction, in accordance with the terms of licences
issued by the Copyright Licensing Agency. Inquiries concerning
reproduction outside those terms should be sent to the publishers.

SAGE Publications Ltd
6 Bonhill Street
London EC2A 4PU

SAGE Publications Inc.
2455 Teller Road
Thousand Oaks, California 91320

SAGE Publications India Pvt Ltd
32, M-Block Market
Greater Kailash – I
New Delhi 110 048

British Library Cataloguing in Publication data
A catalogue record for this book is available
from the British Library
ISBN 0 7619 6067 8
ISBN 0 7619 6068 6 (pbk)

Library of Congress Control Number 2001 132949

Typeset by SIVA Math Setters, Chennai, India
Printed in Great Britain by The Cromwell Press Ltd, Trowbridge, Wiltshire


Notes on Contributors vii
Acknowledgements xiii
Introduction: Transformation in Principles and Practice 1

Part 1 Putting the Practice into Theory 15
1 Institutional Ethnography 17
Dorothy E. Smith

2 Critical Realist Ethnography 53
Sam Porter

3 Framing the Rational in Fieldwork 73
Peter K. Manning

4 Analysing Interaction: Video, Ethnography
and Situated Conduct 99
Christian Heath and Jon Hindmarsh

Part 2 Generalization, Interpretation and Analysis 123
5 Generalization in Interpretive Research 125
Malcolm Williams

6 Representation, Responsibility and Reliability in
Participant-Observation 144
Martín Sánchez-Jankowski

7 Automating the Ineffable: Qualitative Software and the
Meaning of Qualitative Research 161
Nigel G. Fielding

8 Subjectivity and Qualitative Method 179
Valerie Walkerdine, Helen Lucey and June Melody

Part 3 Choices in Context 197
9 Observation and Interviewing: Options and Choices
in Qualitative Research 199
Kathleen Gerson and Ruth Horowitz


10 Qualitative Interviewing: Asking, Listening and Interpreting 225
Jennifer Mason

11 Narrative in Social Research 242
Steph Lawler

Part 4 Power, Participation and Expertise 259

12 Engagement and Evaluation in Qualitative Inquiry 261
Linda McKie

13 Negotiating Power and Expertise in the Field 286
Lynne Haney

14 On Relations between Black Female Researchers
and Participants 300
Tracey Reynolds

Part 5 Reflexivity, the Self and Positioning 311
15 Ethnography and Self: Reflections and Representations 313
Amanda Coffey

16 Reflexivity and the Politics of Qualitative Research 332
Lisa Adkins

17 Techniques for Telling the Reflexive Self 349
Beverley Skeggs

18 Emotions, Fieldwork and Professional Lives 375
Sherryl Kleinman
Index 395


Lisa Adkins is Lecturer in Sociology at the University of Manchester. Her
research interests are in the areas of social theory, feminist theory, sexuality,
gender and economy. Her publications include Gendered Work (Open
University Press, 1995), Sexualizing the Social (with Vicki Merchant,
Macmillan, 1996) and Sex, Sensibility and the Gendered Body (with Janet
Holland, Macmillan, 1996). Recent articles have been published in Theory,
Culture and Society and Economy and Society. She is currently completing a
book called Revisions: Towards a Feminist Sociology of Late Modernity (Open
University Press) which considers theories of identity transformation in
relation to gender and sexuality.

Amanda Coffey is based at the School of Social Sciences, Cardiff University.
She is one of the founding editors of the journal Qualitative Research and a
co-editor of The Handbook of Ethnography (Sage, 2001). Her research interests
include ethnographic representation and (auto)biography in qualitative
method. Her publications include (with Paul Atkinson) Making Sense of
Qualitative Data (Sage, 1996) and The Ethnographic Self (Sage, 1999).

Nigel G. Fielding is Professor of Sociology and Co-Director of the Institute
of Social Research at the University of Surrey. He has taught field methods
and criminology at Surrey since 1978. His research interests are in qualitative
methods, new research technologies and criminal justice. He was editor of
the Howard Journal of Criminal Justice from 1985 to 1998, and is co-editor of
the series ‘New Technologies for Social Research’ (Sage). He has published
twelve books and is currently working on the second edition of Computer
Programs for Qualitative Data Analysis (with E. Weitzman and R. Lee) and a
four-volume set on Interviewing (both for Sage).

Kathleen Gerson is Professor (and Chair) of Sociology at New York
University, where she specializes in the study of gender, work, family and
social change processes. She is the author of several books, including
No Man’s Land: Men’s Changing Commitments to Family and Work (Basic Books,
1993) and Hard Choices: How Women Decide about Work, Career, and Motherhood
(University of California Press, 1985). She is currently at work on a study of
‘Children of the Gender Revolution’, in which she is examining the experi-
ences, responses and gender strategies of the young women and men who
have grown up in non-traditional families over the last few decades of tumul-
tuous work and family changes in American society. Additional ongoing

He is also co-editor of Workplace Studies (with Paul Luff and Christian Heath. Interaction and Technology Research Group. She is currently working on a project about public partici- pation on state medical licensing and disciplinary boards in the United States. Connections. . which lays out a framework for studying globalization ethnographically. Her current research project examines the politics of contemporary welfare reform in the United States through interviews with two generations of low- income women in Maryland and California. Current projects include studies of museums. His prin- cipal interests concern the interactional organization of talk and embodied conduct within workplace settings. Cambridge University Press. Jon Hindmarsh is a Research Fellow in the Work. work arrangements and gender equity in modern workplaces and an examination of how growing conflicts between family and work have created new moral dilemmas that are blurring the gender boundaries between women and men. 1995) won the Charles Horton Cooley Award and Honor and the American Dream: Culture and Identity in a Chicano Community (Rutgers University Press. His recent publications include Technology in Action (with Paul Luff. He has also collaborated with computer scientists on studies of social interaction in virtual reality and is currently extending that work to inform the design of advanced. Cambridge University Press. Ruth Horowitz is Professor of Sociology at New York University and received her PhD from the University of Chicago. such as control centres. He specializes in video-based field studies of social interaction. She is also one of the authors of Global Ethnography: Forces. 1983) received an honourable mention for the C. of operation centres on London Underground and of medical consultations.viii NOTES ON CONTRIBUTORS projects include a study of trends in working time. Her book. Journal of Pragmatics and Symbolic Interaction. These projects also involve the design and deployment of new artefacts and tech- nologies ranging from interactive art installations through to image recogni- tion systems. Inventing the Needy: Gender and the Politics of Welfare in Hungary. 2000). Christian Heath is a Professor at King’s College London and leads the Work. is forthcoming from the University of California Press. Mills Award. and Imaginations in a Postmodern World (University of California Press. the aesthetic experience of art-work and ecological perception of public behaviour. Lynne Haney is an Assistant Professor of Sociology at New York University. anaesthetic rooms and classrooms. King’s College London. 2000). Teen Mothers: Citizens or Dependents? (University of Chicago Press. Recent articles have appeared in the Journal of Contemporary Ethnography.W. She has conducted ethnographic research on the state in both the United States and in Hungary. interactive exhibits in museums and galleries. 2000). and address such varied issues as the embodiment of suffering. Interaction and Technology Research Group in the Management Centre.

He has been a Fellow of Wolfson and Balliol Colleges. He has been a Fellow at the Rockefeller Villa. Subjects (Routledge. Organizing Bodies: Policy. gender and class and relations of expertise. 2000) and Constructing Gendered Bodies (with Kathryn Backett-Milburn. as well as a Visiting Professor at MIT. Her most recent publication is Mothering the Self: Mothers. She is currently planning research into adult and child identities expressed through kinship narratives. He has published some twelve books. Macmillan. numerous chapters and articles in profes- sional journals. Opposing Ambitions: Gender and Identity in an Alternative Organization (University of Chicago Press. at the University of Leeds. in 2000. and involves fieldwork in two large US police departments. 2000). Jennifer Mason is Senior Lecturer in Sociology and Deputy Director of the Centre for Research on Family. Boston. and disclosing domestic violence in primary health care. Wilson Award by the Academy of Criminal Justice Sciences. and Senior Research Associate at the Centre for Socio-Legal Studies. 1996) and co-author of Emotions and Fieldwork (with Martha Copp. childhood and research in sociology and social policy. She is currently writing sociologically informed personal essays for audiences within and beyond academia. He was awarded the Bruce Smith Award and the O. Bellagio. he was Senior Research Scholar at Northeastern College of Criminal Justice in Boston. Her current work includes the development of qualitative evaluation of community health work. Peter K. Chapel Hill.W. 1996) and co-author (with Janet Finch) of Passing On: Kinship and Inheritance in England . Steph Lawler is a Lecturer in Sociology at the University of Durham. Prior to moving to Glasgow in 1999 she undertook sociological research in the Department of General Practice and Primary Care at the University of Aberdeen. 1993). Macmillan. Palgrave. Daughters. She spent several years researching narratives of maternal and daughterly identities and has written on the mother–daughter relationship. NOTES ON CONTRIBUTORS ix Sherryl Kleinman is Professor of Sociology at the University of North Carolina. In 2001. and is named in Who’s Who in America and Who’s Who in the World. Sage. Linda McKie is Research Professor in Sociology at Glasgow Caledonian University. caring and citizen- ship. gender. 2001). Oxford. Institutions and Work (with Nick Watson. She is author of Qualitative Researching (Sage. Kinship and Childhood. She is the author of Equals before God: Seminarians as Humanistic Professionals (University of Chicago Press. Manning is Brooks Chair of Criminal Justice at Northeastern University. His current research is on the rationalization of policing including crime analysis and mapping. Oxford. 1999). where she teaches modules in the sociology of gender. The University at Albany and York University. 1984). Toronto. Recent publi- cations include Gender Power and the Household (with Sophia Bowlby and Susan Gregory. and has just completed Picturing Policing (2001).

London. 2001). Queen’s University Belfast. 1991) and his recent publi- cations include ‘African American poverty and the dispersal of the working class’ (1999) and ‘Using computers to analyze ethnographic field data’ (1998). He is finishing a book on the dynamics of social change and mainte- nance in long-term poverty neighbourhoods and has begun a multi-year research project on violence in inner-city schools in the United States. 2000). philosophy and social research (with Malcolm Williams. Tracey Reynolds is Research Fellow at the Race and Ethnicity Research Unit. 1996) and Negotiating Family Responsibilities (Routledge. social theory (1996). 2001) and the third edition of Social Research (Open University Press. Among other projects. with Zygmunt Bauman. 1998). he has authored and co-edited books on organizational change (1991). 1995) and Transformations: Thinking through Feminism (with colleagues. and Black community activism. Martín Sánchez-Jankowski is Professor of Sociology and Director of the Center for Urban Ethnography at the University of California. Sam Porter is Professor of Nursing Research in the School of Nursing and Midwifery. She also lectures in social research skills at the university. mothers and paid work. 2nd edn 1997). Formerly. Inheritance and Families (Clarendon Press. His latest book is Social Theory and Nursing Practice (Macmillan. His speciality is the sociology of poverty and he has primarily used participant- observation methodology for his research since 1975. social theory and the connection between these two areas. 2000). More recently he has written. 1993). 1996). In addition to articles on the same subjects. social theory and methodology (with Malcolm Williams. he held a post in the Department of Sociology in Queen’s and his research activities span the disciplines of nursing and sociology. he is currently writing a book on reflexivity (Sage) and continuing to edit an international book series (Issues in Society. Thinking Sociologically (Blackwell. As well as engaging in substantive qualitative research. Berkeley. He is the author of Islands in the Street (University of California Press. having previously worked at Durham (1995–99) and Plymouth (1989–95). Her PhD doctoral thesis focused on Black mothering in Britain. Open University Press). work with offenders (1996) and philosophy. Feminist Cultural Theory (Manchester University Press. social research (1993. ethno- graphy (1993). Wills. Routledge. She has also . She was Co-Director of Women’s Studies at Lancaster University 1994–97. South Bank University.x NOTES ON CONTRIBUTORS (Routledge. Tim May is Professor of Sociology at the University of Salford. he has written extensively on research methodology. Her current research interests and publications are around the areas of the Black family. She has published Formations of Class and Gender (Sage. 1997). 1998). Beverley Skeggs is Professor of Sociology at Manchester University.

King’s College London. Valerie is presently Foundation Professor of Critical Psychology. UK. Dorothy E. Theory and Investigations (University of Toronto Press. Helen Lucey and June Melody worked together as the research team for the ESRC research project ‘Transition to Womanhood in 1990s Britain’ at Goldsmiths College. 1990) and Writing the Social: Critique. 2001). Facts. 1987). 1998). Also (with Tim May) he is co-editor of Knowing the Social World (Open University Press. music and sexuality. Routledge. 1999). Helen is Research Fellow in the School of Education. housing need and counter-urbanization. She has been preoccupied since the late 1970s or so with developing the implications for sociology of taking women’s stand- point. 2000) and Introduction to Philosophy of Social Research (with Tim May. and Femininity (Routledge. 1990). The Conceptual Practice of Power (University of Toronto Press. Sexuality and Space. University of London. 1996). Text. and June Melody is training to be a psychoanalyst in London. Smith is in the Department of Sociology and Equity Studies of the University of Toronto. NOTES ON CONTRIBUTORS xi written on postmodernism. Head of the Centre for Critical Psychology and regional head of the School of Psychology at the University of Western Sydney in Australia. Malcolm Williams is Principal Lecturer in Sociology at the University of Plymouth. He is author of Science and Social Science (Routledge. Her most recent research project is on Violence. Her books include The Everyday World as Problematic: A Feminist Sociology (University of Toronto Press. . Their latest book is Growing Up Girl: Psychosocial Explorations of Gender and Class (Palgrave. Valerie Walkerdine. He has also published a number of articles in the areas of research methodology. toilets.


ACKNOWLEDGEMENTS First. They put up with my constant requests and even my ‘nudges’ to get chapters completed on time and were also kind enough to comment on my introductory chapter. my love and gratitude. Zoe Elliott and the editorial and production teams. Cian and Calum. Those at Sage Publications were very supportive of the project and thanks are due to Michael Carmichael. My thanks to those friends who have remained so supportive and to Dee. . my thanks to all of the contributors. as well as to Malcolm Reed for compiling the index and Christine Firth for performing such excellent work on the manuscript.


along with the need to share experiences. critical realism. critical theory. thinking about the practice of qualitative research. Before moving on to provide an overview of the chapters. PRACTICE. Although mostly aimed at positivism and empiricism. feminisms. postmodernism and post-structuralism – is not my task here. social constructionist perspectives. methodology and method. The result is a book structured around ‘issues in practice’ as its main focus. Overall. it is to draw out some of the themes in order to provide a context for the chapters and part headings that appear in this volume. for instance the relationships between philosophy. To detail these transfor- mations. as well as pleasurable and supportive. theory. in various ways. not collected. or as a result of the presuppositions . but also to permit chapters to be exchanged and commented upon. In the planning of this volume a particular process was adopted. brought about by a number of different traditions – for example.INTRODUCTION: Transformation in principles and practice Tim May Welcome to Qualitative Research in Action. the purpose was to enable a the- matic coherence to emerge within the volume as a whole. Given this. Nevertheless. It was also recog- nized that while we talk about the links between process and product in research practice. it is necessary to situate them in terms of the issues that have informed. Whether overtly. this is often not the case when it comes to the production of edited collections. and it is the process of production that is fundamentally related to the product (May 2001). The reasons for this were not only due to the geographical spread of the contributors. the idea of ‘pairing’ was introduced. We can now observe that data are produced. it has resulted in extensions of particular discourses into terrains that were once presumed to be clear in their demar- cation points. PROCESS AND TEXTS We have witnessed in the latter part of the twentieth century a number of critiques concerning the status of social research in society. The aim here was to make the process of writing for an edited collection more thematic.

These include the desirability of various courses of actions. the decisions that are made about theory. fiction and art. poetry. originally published in 1959. while Barry Smart (1993) finds references to the term in 1930s lite- rary criticism. Those napping in the cosy slumbers of past scientific ‘pretensions’ can then find their practices being characterized as branches of literary criticism. in reaction to these criticisms. Here he noted how social science had inherited terms which. The latter. In the process scientism – defined as the belief that science is the only form of legitimate knowledge – becomes confused with science. In The Sociological Imagination. As a result we find calls for a return to a ‘scientific basis’ for disciplines mixing with those who denounce social research in favour of other mediums of representation. If the epoch which has given us these critiques. claim to speak in the name of a separate and unproblematic reality should be exposed to scrutiny. While the idea that one can. is concerned with the search for new values. irrelevant. as well as recognition of different forms of life. There are many other examples that could have been used to illustrate this point. Particular ideas of neutrality. added emphasis). can be moulded by considerations that lie beyond the confines of its boundaries. ‘become unwieldy. although outdated. Here we can detect movements in opposite directions and this detracts from the productive potential that comes with engagement. We see the same tenden- cies in both claims: that is. if generalized to contem- porary situations. ethics and politics are now open to routine scrutiny. Mills had a chapter on ‘reason and freedom’. without question. however. if mixed with a sensitivity to context and a willingness to engage in an under- standing of the relationship between justification and application that is not taken to be beyond question. remain rooted in practice. Polarizations between an unproblematic science and acts of literary deconstruction do little to aid understanding. . Perhaps we may call it: The Fourth Epoch’ (Mills 1970 [1959]: 184. He then moved on to argue that these ‘standard categories of thought’. for example. to legislate over the constitution and nature of social reality.2 INTRODUCTION that are inevitably embedded within ways of thinking that inform practice and so often remain beyond question. from those who argue that nothing has changed to those for whom anything less than total embrace is an act of betrayal. so too should the claim to speak in the name of different realities as mediated by alternative modes of representation. identities and ways of life. not convincing … so now The Modern Age is being succeeded by a post-modern period. The knower (as researcher) is now implicated in the construction of the known (the dynamics and content of society and social relations). methodology. At this point we might note that this trend is nothing new. such as the maintenance of objectivity through positioning the researcher as nothing but a passive instrument of data collection. per- haps it is not surprising that we can range. Jürgen Habermas (1992) argues that the opening for postmodernism in western thought began in the late nineteenth century in the writings of Nietzsche. methods. are now exposed as falsehoods that seek to mask the realities of the research process.

cultural capital within academic fields. With the struggle for academic capital in place. A series of acts of demolition on so-called classic texts can assist in the process of accumu- lating. The effect is to produce competing academic camps with those defending established procedures pitted against those who set themselves the task of finding tacit assumptions that represent nothing more than nostalgic yearn- ings for a bygone age. for the purposes of illuminating the dynamics of social issues. INTRODUCTION 3 There is also the transmission and effect of these critiques as they become aligned with modes of scholastic communication to consider. In the face of these trends it is possible to form the distinct impression that methodological discussions are now trapped within descending interpreta- tive circles. The result is that rep- resentations can be rendered so incoherent that engagement is difficult. Yet how are the interpreters authorized to make such interpretations in the first place? Would this be something to do with the institutional authority that is bestowed upon them . These texts may be laid bare in order to render public just how they cannot live up to their own aspirations. does such activity become counter-productive from the point of view of research practice and understanding itself? When does it cease to inform in order to change prac- tice and instead undermine it and so lead to paralysis and inactivity? Take. appear as more subtle and useful to the present age than previous interpretations might have suggested. At what point. let us activate the differences’ (Lyotard 1993: 46). The search for true knowledge is the target of attack for representing the ‘fantasy to seize reality’ and thus the solution becomes: ‘Let us wage war on totality. let alone those standards which those who charge themselves with this undertaking invoke. We end up with accounts that can reproduce the very targets of Derrida’s cri- tiques: the closure of texts and the centrality of the subject in the production of those texts. however. let us be witnesses to the unpresentable. We then await the next stage for those thinkers subject to acts of deconstruction: the process of resurrection. Careers are then forged in acts of deconstruction against which another set in the intellectual class can then sharpen their pencils. methodological translations of the works of Jacques Derrida (1978). what then emerges is a whole new industry for textual reflections on the futility of this enterprise (May 1998). paradoxically. reproduce the very ego-identity that is the subject of his critical interventions (see Norris 1987). At this point those same thinkers. The project is to expose the presuppositions of those who were once canonized as represen- tatives of a tradition that budding apprentices are expected to emulate in their practice. when translated in different ways accord- ing to different influences. These can lead to reflexive accounts that. Such activity is a vital part of the vibrancy of intellectual work and func- tions as an instrument for sensitizing researchers to the consequences of practices and assumptions. reputations may be forged via interpretations of interpretations. for example. if not impossible. to deploy the tools of analysis of Pierre Bourdieu (Bourdieu and Wacquant 1992). Manifest in confessions of failed attempts to discover an unproblematic reality.

theoretically informed empirical work. For Foucault. after all. contrary to the attitude of those who regard philo- sophical matters as detracting from the ‘real’ work. Built into assumptions are epistemological and ontological presuppositions that render the world intelligible.4 INTRODUCTION and that enables a distance to be maintained from the practice and products of systematic social investigation? Of course there are clear advantages of such distance. but also how we dominate ourselves. it is because researchers have a distance from the necessity they examine in order to turn practice into an object of investigation. From a more critical genre. to assume that the end of critique is some consensual state of affairs runs the danger of the constitution of a complacent attitude that. particularly when compared with those whose practice is driven by the interests of sponsors who make specific demands according to the pursuit of narrow interests. disguises non-consensuality in a celebration of finality (Foucault 1984). What is at stake is not just how others dominate us. by default. if a certain form of practice and contemplation – as that which enables research to be undertaken in the first place – is born in a double movement of reflexivity in terms of possessing a point of view on the point of view (Bourdieu and Wacquant 1992). We find the same impulse in feminist-inspired research where reconstruction is identified alongside the need for deconstruction in order to remain sensitive to the working assumptions that inform practice (Harding and Hintikka 1983). this provides for the possibility of things being other than they are via a critique of what Roberto Unger (1987) has called ‘false necessity’. but not suffi- cient condition. to guard against ‘power as evil’. must be . we find Pierre Bourdieu’s (2000) concern to conduct research as a contribution to a ‘realpolitik of reason’. The potential for a critical approach to research work. consid- ered genealogically. This same genre can be captured in Foucault’s work. Add to this the fact that the world is not about to stop and listen to such debates and this tells us something about the ways in which our domi- nant modes of organizing social relations take their revenge on discourses. If we turn our attention to such matters we might then be able to expose what Michel Foucault (1991) called the ‘limits of appropriation’ of discourse. In this way we may not see it as all encompassing of social relations and begin to ask how it is that the analyst may claim a privileged vantagepoint in the face of what are taken to be overwhelming social forces? The issue now turns to reconstruction. Similarly. The limits to reflexivity then inhere in a willingness to subject one’s own position and what it does or does not authorize as a result. A critical ethos must be seen as a necessary. irrelevant. Practice. to critical scrutiny (May 1999. However. can be conducted in order to illuminate issues and bring to the attention of a wider audience the dynamics and consequences of social relations. This appears as a process without end. 2000). in order to avoid the traps of a ‘false’ universalism and particular- ism. Yet it can so easily lapse into an indulgence that misses its mark. These issues are not. is its ability not only to ‘reveal reality’. Engaged. but also to ’deconstruct necessity’ (Hoy 1998).

Different forms of representation are then assumed to take precedence over the results of systematic social investigation and in so doing simply reverse the scales. rather than having the potential to be complemen- tary in its discovery. who are always more than willing to supply that need’ . but not necessar- ily the next generation of academics who may now be doctoral students and those on temporary research contracts. The study of the conditions of knowledge production in terms of the positioning of the researcher is thus a guard against not only the conflation of scholastic with practical reason. the potential for further illumination of social dynamics and situations can so easily slip from being complemen- tary. politicians. continual and conscious monitoring of actions within actions would bring them to a halt. forms of representation aside from social research from having a regulatory and reflexive function. These are the armies of ‘journalists. by default. Further. Yet questions will still be begged when we ask of such forms of representation: whom do they affect? Under what circumstances and for what reasons? Utilizing what resources and with what overall consequences? Systematic social investigation enables us to seek answers to such questions. We also have to ask another question: what are the costs of these confla- tions to future generations of researchers? What about those who are under pressure to produce results according to the edicts of institutions and spon- sors who. pundits. Although language carries a degree of reflexivity (Garfinkel 1967. and pop-theorists. Fiction becomes reality. It then moves. rather than the results of systematic social investigation. but also the complacency and arrogance that is born in dispositions that have the potential to accom- pany all modes of thought – whatever their claims may be concerning the constitution of social reality. Harré 1998). Without a gaze turned towards the continuing endeavours and positions of researchers. of course. after all. that it is often the art of persuasion. INTRODUCTION 5 indifferent to the conditions of its possibility in order to be practice as such. what may be omitted from such deliberations are the conditions that enable and constrain actions. Similarly. how can the policy process be informed and changed while noting. pay their wages? In the face of this how can critical inter- ventions provide those researchers with the resources to challenge organizations that will continue to commission research? It is they who are positioned to completely ignore such critical interventions. If there is a withdrawal from fields of current endeavour there are plenty of those who are unfettered by the latest critiques concerning the quest to uncover realities that are only too happy to fill it. however. Imagine the paralysis that would come with repeated reflections upon every act prior and subsequent to its execution. to a constitutive one. to having a substitutory role in relation to research practice. which informs such outcomes? Here we can detect another possible consequence in relation to engage- ments between social research and social life. while experimenting with new forms of writing. those who once sought something called an unproblematic ‘reality’ can turn their attention to past mistakes. Again.

Along with these we have witnessed the introduction of new technologies of data analysis. ‘instrumental positivism’ (Bryant 1985) still appears to find a way of marching onwards. narratives. As a result we learn about what we may and may not expect of the product. None of this is to suggest that neatly demarcated boundaries exist. under what circumstances and with what effects. representation and generaliza- tion. CONTRIBUTIONS: ISSUES. processes and practices of research. In what is taken to be a methodologically post-positivist/empiricist/modernist age. to those who have recently embarked upon their careers. issues thereby emerge from within these chapters. reflexivity. The contributors are reflecting upon their experiences in terms of what it has informed them about the process of conducting qualitative research. Important. Different ways of organi- zing research can also open up new possibilities (Gulbenkian Commission 1996). but often-neglected. emo- tions. What emerges is a sense of the issues that continue to arise in practice. from other collections. Those who wish to change practice. the book is divided into five parts. These include the relations between fieldwork and social identity. In the process new terrains of inquiry have been opened up. In the productive agonisms that can and should exist between social research and social life. THEMES AND CONTENT There is no doubt that there are positive benefits in the processes of rethink- ing the issues. It has been designed with the intention of assisting the process of reconstruction along- side a continued monitoring of the working assumptions of research prac- tice. What distinguishes this volume.6 INTRODUCTION (McCarthy in Hoy and McCarthy 1994: 220). All of them are informed by the theme of interrogating practice in terms of . we need to understand much more about how and under what circumstances it can be deployed without a capitulation to the power of those social forces that seek to mould its practices and findings in their name. leads to a limited understanding of its place and value in social life. participation. as well as the actual content and context of its prac- tice. but do not take this into account. They come from different intellectual traditions and range from those well established in their disciplines. evaluation research and ways of combining methods to enhance insight into the dynamics of social life. but also the contributors themselves. Yet it is upon qualitative research that these critiques have so often alighted and it is for these reasons that this volume was brought together to consider their implications for actual practice. actions. For the purposes of assisting the reader. or subjected to scrutiny in new ways. It is to say that a failure to understand the forces which act upon the process of social research and the conditions under which it is enacted. All of these topics and more are examined in this collection. as well as how they are addressed. is not only the focus and scope of its contributions. simply miss the target.

as noted earlier. the self and positioning. participation and expertise in the research process. is explicitly focused around this general issue. all of which takes place in an environment which is supposed to reflect ‘laws’ of supply and demand. may often be overlooked. Part 2 is then organized around concerns with generalization. they are distinctive and so it is hoped that the reader. Part 4 specifically focuses upon issues of power. is to map how social relations are coordinated by and contribute to activities in different sites in order to bring to people’s attention how their lives and actions relate to those of others. but also how it addresses the relations between the conduct of research and those who are its subjects and co-producers. bearing in mind the transformations that have taken place in thinking about the role and process of research in understanding social relations. however. will be better able to engage with issues in practice that. The first part. as it is. once equipped with these accounts. interpretation and analysis and Part 3 examines methodological choices in practice. In the process we gain an understanding of not only the main tenets of this approach in relation to its aims and modes of engagement with the social world. she considers the relations between material conditions. In Chapter 2 Sam Porter discusses the extent to which qualitative research can be used to examine the relationship between social structures and social actions. INTRODUCTION 7 its potential for explaining social relations via actual research examples. The purpose of the analyst. . dis- courses and school. Beginning with an overview of two classic texts – Durkheim’s Elementary Forms of Religious Life and Weber’s Protestant Ethic and the Spirit of Capitalism – he argues that Durkheim did not attempt to provide a methodo- logical justification for his approach. she argues. This provides him with an opening for an exposition of critical realism via examples drawn from his own work. Claims to explanation often rest upon the invoking of rational action of some type. The final part is organized around the themes of reflexivity. Using examples drawn from work with collaborators. He then moves on to trace the development of qualitative work. home and work. Smith. within the women’s movement and focused upon the ‘everyday world as problematic’. about which they may not be aware. while Weber’s idea of verstehen was somewhat vague. In this chapter we find a detailed overview of her approach to studying social relations by taking a procedural mode of understanding to what has become a highly influential practice rooted. PUTTING THE PRACTICE INTO THEORY Part 1 opens with a contribution by Dorothy E. Nevertheless. via phenomenology and postmodernist approaches and in so doing notes a gradual drift away from an examination of social structures. These parts cannot be exhaustive nor can there be any claim of this type. In the west we live in a system of organizing social relations that is presumed to be underpinned by actions that aim to maximize individual utility through the selection of particular means.

this combines with a context-sensitive rationality that also contains what might otherwise be termed ‘irrational’ behaviour. ratio- nality. finds its expression in the work of Harold Garfinkel (1967). Garfinkel’s approach is to start from the analysis of local ‘experience structures’ and not to generalize from social structures to per- sonalities. for example. Yet it is seen in these terms only if the local settings in which human action takes place are ignored. he examines a number of . Ethnomethodology was born in this move.8 INTRODUCTION As this is the ideological basis of a system that is so taken-for-granted in the habitual actions that inform everyday life. Nevertheless. but also renewing activities. the interactions between people and technology. While explicitly seek- ing to avoid a quantitative–qualitative divide. what we find within police work is a covering. INTERPRETATION AND ANALYSIS In the opening chapter to Part 2. this is an issue that requires empirical investigation. While any identified features of actions may be generalizable. Indeed. Malcolm Williams picks up on the issue of generalization that Sam Porter raises in Chapter 2. including the post- modern insight that the grounds of human action are often irrational. as opposed to one that is imposed by the observer. A concern with a context-sensitive rationality. it would appear to be the domi- nant way of viewing human action. the authors argue that the issues which arise in research between the subject and object and talk and non-verbal communication. they argue. As a result researchers have to be aware of the assumptions they make in understanding the actions of those they study. that is the subject of Chapter 4 by Christian Heath and Jon Hindmarsh. The authors then move on to demonstrate how conversation analysis and ethnomethodology can provide data that is more usually asso- ciated with ethnographic work and this enables a greater understanding of. that enables the modes of sense-making in local settings and the production of intelligibility to be understood. this story is a long way from an accurate description of human actions. when it comes to utilizing such ideas in research. to produce a social scene. In the course of his account he takes various approaches to rationality to task. However. on the part of social actors. Using video and conversation data. GENERALIZATION. are amenable to study from an ethnomethodological per- spective. It is this perspective. there is a frequent conflation of this norma- tive background with the presumed neutrality of its modes of description. Taking themes from transformations in social thought. or procedural. Peter K. They present studies to demonstrate how there are continual and concerted efforts. in terms of the study of institutional talk and its relationship to ethnography. It is an analysis of these. These are not only context-sensitive and producing actions. that is orientated to working in a context where legal rules have to be taken into consideration. Despite this. Manning (Chapter 3) engages with these issues via examples taken from his work on police detectives.

that it is only survey work which is reliable. At the same time. This is regarded as being able not only to take on board the issues that he raises and so account for the claims often made by qualitative researchers. we must also add the matter of how observed actions should be reported. this is associated with the issues of confidentiality. Originally designed for the purpose of data analysis. . he employs moderatum generalizations within a pluralist approach to research practice. This history. it is often the case that as long as the researcher can somehow become ‘detached’ from the production of their findings. In qualitative research matters relating to selec- tivity for the purpose of producing accurate representations were to receive a boost with the introduction of dedicated computer software. sequences of thought. including not only generalization. often reproduced by qualitative researchers themselves. In terms of a practical-political problem of represen- tation. action and meaning and how many supporting data need to be included in field reports. Chapter 6 by Martín Sánchez-Jankowski seeks some practical solutions to these core issues. while noting in Chapter 7 that ‘epistemological preoccu- pations are more enduring than any technology’. In a wide-ranging discussion he notes how errors may be reduced in observation research through attention to a set of core issues. but also to provide for a politically informed engagement which is necessary for social transformation. By examining the sequencing of actions there are ways in which an account can represent facets of observed action. procedures and principles of quali- tative research itself. To these questions. This also has the benefit of addressing the prejudice. he argues that there are clearly ways in which qualitative research seeks to generalize from case studies to other circumstances. its use has now extended to encompass collection. Nigel G. arithmetical and graphical forms of representation still hold power over an audience as if the process of selec- tivity had not taken place. examines the history of different packages and their relationship to the qualitative research process. INTRODUCTION 9 epistemological and ontological themes. Now add issues of generalization to one of the most important issues that researchers face: the movement from fieldwork to writing up. Thus. Taking case studies to illustrate his points. underpinned by the story of code-and-retrieve. While quali- tative work seeks to represent social processes. then provides a basis for him to move on to demystify these technological changes and examine their effects upon the practices. often by invok- ing the notion of ‘typicality’. its validity and reliability is assumed to be greater. ‘Solutions’ to methodological and theoretical issues often appear in the form of ‘techno-fixes’. Seeking to steer a course between ideographic and nomothetic approaches. ‘what you see is what you get!’ Nevertheless. this does not relieve the researcher of some core responsibilities. however. but also reliability. A consideration of these matters often derives from a phrase that is indicative of the natural- istic underpinnings of observation-based research: that is. Fielding. we often read accounts and ask of the observer how did they know that? Particularly when it comes to unobserved phenomena. literature searches and the writing-up process itself.

subjectivity and narratives. in conceptual and epistemo- logical terms. Using examples from their work. Kathleen Gerson and Ruth Horowitz con- sider the relations between interviewing and observation-based studies. By asking what interviews are and what they do. Jennifer Mason’s chapter raises issues associated with time. Although often taken to be methods with different epistemological assump- tions and theoretically at odds with one another. how some aspects of the ‘social’ cannot be captured through a concentration on talk alone. What we end up with. to consider how people live with the contradictions and demands that are placed upon their everyday lives. A decontextualized form of knowledge gather- ing thus becomes highly problematic. during the course of the discussion. when combined in practice. Valerie Walkerdine. To overcome this tendency she argues that the interview should be seen as a process of co-participation in which both parties regard it as a site of knowledge production. enrich our insights into social life. They note that observational methods provide information concerning how individuals and groups behave in a range of social settings. there is also the place of the researcher in the process. Helen Lucey and June Melody (Chapter 8) examine fantasy. Just what is the significance of narratives in terms of how they are deployed in social life in order to construct accounts and social identities? It is these types of questions that Steph Lawler raises in Chapter 11 by taking narratives as accounts that bring together past and present with self and . Manning. therefore. At this point interviewers are often faced with the act of interviewing being a static-causal snapshot when they actually seek to understand social processes. CHOICES IN CONTEXT With choices in context in mind. they are also seen to com- plement each other in significant ways. while interviews uncover the percep- tions. but also production. is a demonstration of how what are often maintained to be opposing positions can. there is an examination of subjectivity and its rela- tion to the production of fieldnotes and how differing subjectivities inform interpretations of the same data.10 INTRODUCTION Aside from the role of technology in analysis. To elaborate upon this idea she considers it in terms of her own work on families and kinship noting. motives and accounts that people offer for their actions and beliefs. without necessarily abandoning the insights that they have generated. she is then in a position to explore issues associated with generalization. We thus return to an earlier theme raised by Peter K. With this in mind. Jennifer Mason continues in Chapter 10 with the theme of choice in methods. transference and counter- transference in relation to not only data analysis. only this time focused upon a different question: how is it that a non-rational understanding of the actions of researchers themselves affects the fieldwork process and product? The concern here is to move beyond ideas of narratives and discourses. Chapter 9 compares and contrasts these two main research techniques via examples drawn from their own work.

state institutions are viewed as spaces in which ‘relations of the ruling’ are contested among women who. to borrow a characterization from the work of Paul Ricoeur (1994). within an overall plot. Linda McKie notes that it is often assumed qualitative fieldwork enables the voices of respondents to be heard in ways that quantitative work does not permit. As a result these condi- tions create particular issues for reflexive researchers. exercise power over women as clients. in the opening chapter to Part 4. as state actors. pro- vide critical cases for viewing the dynamics of the fieldwork process. Smith. Add to this the flow of power and its sites of production and this will have an effect upon how researchers can engage with communities and other stakeholders in the research process. Following the work of Dorothy E. her own research practice did not then turn inwards. She then turns to how a blurring between ‘rulers’ and ‘ruled’ occurred in terms of the methodological injunction to listen to the ‘voices of women’. Thus. be said. In order to consider how and under what circumstances these arise and the manner in which they may be acted upon. It is this theme that informs the next two chapters. but outwards via an analysis of structural forces and ‘ethnohistory’. conduct and dissemination of the research itself. along with actions and characters. participation and expertise are clearly matters that inform the con- duct of research in the field. However. These are expressed in terms of transformations over time. Drawing upon her own experiences in two evaluation pro- jects in Scotland. she highlights the importance of dia- logue and deliberation. she discusses feminist-inspired debates about co-participation. By considering her own experiences in these contexts and how the projects unfolded and what actions were taken during the process. Mix this with evaluation studies and it can raise expectations among partici- pants that may not be met by the process and its product. INTRODUCTION 11 other. is to steer a middle course between Cartesian ‘epistemic exhaltations’ of the self and its apparent ‘humiliation’ in the hands of Nietzsche. These studies. she then notes the tensions between informing the policy process in order to improve conditions for local people and the generation of information that enhances policy-makers’ ‘power over’ communities. . Lynne Haney (Chapter 13) examines the roles of power and negotiation in the context of two ethnographic studies – one in California and the other in Hungary – she has conducted on the state. she argues. POWER. What is then required on the part of the researcher is a heightened sense of the dynamics of power in terms of how they inform the design. To understand narratives it is necessary to situate people within particular historical and cultural milieus in order to see how they are indicative of what may and what may not. Linda McKie provides us with a very good illustration of the practical issues that arise in the conduct of evaluation work and how this affects its credibility and potential within different settings. The result. PARTICIPATION AND EXPERTISE Power.

to automatically assume an imbalance of power in favour of the researcher is highly problem- atic when it comes to the dynamics of the fieldwork process itself. in terms of their identity. she argues that the turn towards particular ideas on reflexivity in social research entails relations between the knower and known that permit only certain voices to be heard. generalization. The former refers to an agent’s ability to monitor and account for their actions. this has sensitized us to important issues for practice.12 INTRODUCTION Lynne Haney regards an inward-looking practice as one that focuses upon the position of the researcher without due regard to wider social processes. she produces an account which questions the so-called reflexive turn in terms of its ability to amplify marginalized voices. Beverley Skeggs picks up the idea of mobile selves in Chapter 17. This does not sug- gest a balance of power. but also gender and class. sexuality and work. REFLEXIVITY. Tracey Reynolds (Chapter 14) brings this to her account of being a black female researcher interviewing other black women in contemporary Britain. she argues that power should be seen as shift- ing and renegotiating itself according to differing contexts in terms not only of race. Lisa Adkins then moves our focus to the politics of reflexivity in Chapter 16. This raises key questions: what kind of self is required to be a reflexive researcher? What kind of narratives do research subjects/ co-producers need to perform in order to be reflexive? In asking such ques- tions. An understanding of social issues is thereby in dan- ger of being abandoned in favour of introspection without engagement. analysis and power in fieldwork. Such thinking may be argued to have informed debates concerning the posi- tion of white female researchers researching black women. but how does this relate to the idea of the reflexive turn in social research. THE SELF AND POSITIONING Contributors have sought to illuminate issues surrounding representation. while the latter is concerned with the . as I noted ear- lier. By taking a relational approach. the self and positioning? The final part thus starts with Amanda Coffey considering the literary turn in ethnographic writings and its relationship to the self. From this point of view. Amanda Coffey therefore turns our attention to these issues and the limits to autoethnography for the purpose of representation. To this extent the idea of producing selves within research texts should be viewed as only one part of the ways in which relations between the researcher and the social settings they seek to understand should be consid- ered. A vision of the mobility of the knower. While. is underpinned by the assumption that they are able to move across boundaries. simply that the assumption of the powerful researcher and the powerless research subject requires a more nuanced understanding. In the grander claims of social theory there is often a conflation of two dimensions of action: ability and capability. Drawing upon her studies of gender. it has also had the effect of diverting attention away from what is discovered as a result of the research process itself.

under what circumstances and utilizing what resources? In drawing upon her own fieldwork. Without this consideration in place. but for which individualism is a totally false description of the social world. positions and practices. In these circumstances individualism may flourish and the commitment and passion to conduct research can be bracketed. this so easily lapses into a constructivist idealism that misses its mark. INTRODUCTION 13 power to act. as well as an understanding of the conditions of knowledge production itself. Such a belief is frequently perpetuated within the conditions of knowl- edge production itself. In the first instance this neces- sitates an understanding of the issues that arise within research practice. but who. SUMMARY All of the chapters in this collection constitute core insights into the per- spectives. Nevertheless. . experiences and issues that inform and arise from the process and practice of qualitative research. as opposed to a study of practices in relation to position- ing. An assumption is often made that emotions are a block to objective analy- sis. concerns dispositions and positions within fields of relations characterized by the distribution of differing forms of capital (Bourdieu 1992). Sherryl Kleinman (Chapter 18) thus starts with such conditions and their effects upon her iden- tity and understandably uncertainty. on the part of some. This is particularly paradoxical when a discipline certainly respects and celebrates the individual. however. The potential to inform practice may then be derived from an explanation of the relations that exist between dispositions. but also those who are the subjects and co-producers of qualitative research. The result is an insightful account of how it is that emotions can inform not only a greater understanding of ourselves. following the work of Pierre Bourdieu. aloof detachment can mix with the posturing that accompanies positions informed by the accumulation of cultural capital. there is a tendency to regard textual critiques as somehow sufficient for the changing of practice. for example. With this in mind the issues surrounding and informing mobility should not be the assump- tion that all are mobile. Beverley Skeggs argues that attention should be turned to understanding and explaining why some people are not mobile and how their fixed positions are relied upon for the mobility of others. This. she charts how she took such feelings as an impetus for further understanding through her own fieldwork. calls to reflexivity become nothing more than a licence to confess. In university departments. As I noted at the beginning of this introduction. The chapters in this volume are a valuable and insightful contribution to that process. In the course of her discussion she thus uncovers boundaries to poten- tiality whose existence creates a refusal. Without sensitivity to the pressures and experiences that inform research. as Lisa Adkins also argues in Chapter 16. to see reflexivity and mobility as privileges born of positioning.

P. Lyotard. London: Routledge. Reidel. European Journal of Social Theory 3(2): 157–173. translated by R.W. (1967) Studies in Ethnomethodology. http://www. a Work in Constructive Social Theory. P. J. London: D. London: Fontana. London: Harvester Wheatsheaf. D. Burchell. Cambridge: Polity.) Postmodernism: A Reader. (1994) Oneself as Another. P. Oxford: Blackwell. London: Macmillan. T. Unger. Nice. T. Buckingham: Open University Press. belonging and reflexivity’.org. Cambridge: Cambridge University Press. Smart. Mills. T.socresonline. in G. (2000) Pascalian Meditations. Garfinkel. Moss (ed.) The Later Foucault: Politics and Philosophy.J. Blamey. D. Cambridge: Polity. M. Gulbenkian Commission (1996) Open the Social Sciences: Report of the Gulbenkian Commission on the Restructuring of the Social Sciences. (1970 [1959]) The Sociological Imagination. (1993) Postmodernity. (eds) (1983) Discovering Reality: Feminist Perspectives on Epistemology. (1985) Positivism in Social Theory and Research. C. . Foucault. London: Harvester Wheatsheaf. Harré. Foucault. S. L. (1992) The Logic of Practice. (1991) ‘Politics and the study of discourse’. edited by P. Hoy. J. translated by K. (1992) The Philosophical Discourse of Modernity: Twelve Lectures. Harmondsworth: Penguin.B. 3rd edn. H. Derrida. (1978) Writing and Difference. Bryant. Chicago: University of Chicago Press. (1998) ‘Foucault and critical theory’. (1992) An Invitation to Reflexive Sociology.uk/socresonline/4/3/may. (1998) ‘Reflexivity in the age of reconstructive social science’. International Journal of Methodology: Theory and Practice 1(1): 7–24. (1998) The Singular Self. (1994) Critical Theory. London: Sage. R.html May. Bourdieu. (1999) ‘Reflexivity and sociological practice’. Bourdieu. (1984) The Foucault Reader.C. C.14 INTRODUCTION REFERENCES Bourdieu. C. T. (2000) ‘The future of critique? Positioning. (2001) Social Research: Issues. Cambridge: Polity. T. Cambridge: Polity. London: Routledge. Stanford. B. and Hintikka. Gordon and P. Methodology. Ricoeur. Harmondsworth: Penguin. Docherty (ed. R. London: Sage. Methods and Process. (1993) ‘Answering the question: what is postmodernism?’. Norris. and Wacquant. NJ: Prentice- Hall. C. (1987) Derrida. Miller (eds) The Foucault Effect: Studies in Governmentality. Hoy.C. CA: Stanford University Press. in J. M. Metaphysics. May. Englewood Cliffs. Harding. J-F. Habermas. Sociological Research Online 4(3). (1987) Social Theory: Its Situation and its Tasks – A Critical Introduction to Politics. translated by R.M. Rabinow. in T. P. Nice. and Philosophy of Science. and McCarthy. M. May. May.



Here. Smith Note: in this introduction to institutional ethnography as a method of inquiry. I have had in mind a hypertext pro- cedure. The method of inquiry to be described in this chapter originates in the women’s movement of the 1970s in North America. In the women’s movement of the 1970s I learned to take my experience as a woman as foundational to how I could know the world. both of the university and of the extra-local relations of academic discourse. students. neighbours. It participates in a DISCOURSE in which particular others appear only as their printed names in texts. Particularized relationships emerge within institutionalized forms of coordi- nation. read- ers have the equivalent of a button to press to shift to a locale where they will find an account of how that concept is being used. and so on. cues and initiatives. The consciousness that organizes household work and childcare is highly attentive to the particularities of the local setting – the physical layout of the household. The consciousness that organizes and is organized in the university setting and in relation to academic work is entirely different.1 INSTITUTIONAL ETHNOGRAPHY Dorothy E. calling the kids in from play to get ready for school. connecting her or him with others known and unknown in an impersonal organization. I make use of various concepts that organize that method. were radically different modes of consciousness. checking the refrigerator to see what’s there for supper. administrators and others. From a standpoint in the everyday world. the objectified social relations of my work in the uni- versity came into view for me in a new way. The two modes of consciousness cannot coexist. extra-local relations that the university sustains. I discovered then my double life of household/mothering and the university as a daily traverse across the line of fault between a woman’s life in the particularities of home and children and the impersonal. in these two work situations. supervisors. I could see how the institutional . When a concept is encountered that has a specialized use in this context. It is a conscious- ness that coordinates multiple particular details. Here the subject participates in relations that extend beyond the local and particular. In this case the equivalent of a button is a word in upper-case: ‘institutional ETHNOGRAPHY’ signals a concept provided with an explanation in an alphabetized list of such concepts at the end of the chapter. putting clean sheets on the beds. taking in the state of the floors. partner. or positioned as members of definite classes of others – colleagues. involving relationships with particularized others – children.

In remaking sociology. to institutional ethnography.18 PUTTING THE PRACTICE INTO THEORY order of which sociology was part was itself a production in and of people’s everyday activities. the sociological project is one that takes up the everyday world as a problematic for investigation. Though the latter are important. it does not take its problems or questions from one or other variant of sociological dis- course – symbolic interaction. but the central project is one of inquiry which begins with the issues and problems of people’s lives and develops inquiry from the standpoint of their experience in and of the actualities of their everyday living. This doesn’t mean that it makes no use of such theories. Marxism. In contrast to other sociologies. I wanted to remake sociology from the ground up so that. organized by and contributes to social relations coordinating activities in multiple local sites. confined to description of local social organization or to expressions of people’s own experiences. Starting to rediscover the social from the standpoint in the everyday world of our experience was essential to a critique of the language of sociological discourse. indeed essential. Our directly known worlds are not self-contained or self-explicating despite the intimacy of our knowledge of them. but that it connected people translocally across multiple local settings. Every local setting of people’s activity is permeated. rather than the object being to explain people’s social behaviour. The sociology I had been trained in was written almost exclu- sively by men from their viewpoint. feminists evolved a critique from that basis and also sought to remake the discipline to enable the experiential to be spoken with authority. ethnomethodology or other ‘school’ of sociological thinking and research. however. The work of the sociologist is to discover these relations and to map them so that people can begin to see how their own lives and work are hooked into the lives and work of others in relations of which most of us are not aware. write and teach and which claimed objectivity was deeply infected with assumptions that relied on excluding women and their concerns and experience from the discourse. we run into the problem that we cannot grasp how they are put together from within them as they are experienced. INSTITUTIONAL ETHNOGRAPHY The sociology that has come out of that experience has come to be called ‘INSTITUTIONal ethnography’. It is not. If we take the idea of being in people’s everyday/everynight worlds seriously. I called this ‘making the everyday world a sociological problematic’ (Smith 1979. My own work was part of this movement with the discourse (Smith 1974). in which our EVERYDAY/EVERYNIGHT WORLDS would be rediscovered as they are organized by social relations not wholly visible within them. 1987). The pronouns ‘he’ and ‘him’ were treated as the universal subject. The everyday/everynight of our contemporary living is . the discipline could be turned upside down to become a sociology for women (Smith 1977). The women’s movement in sociology was slowly learning how to recognize the extent to which the sociology in which we had learned to talk.

It is this relationship that cre- ates the ambiguities of the power relationships that Tracey Reynolds (Chapter 14 in this volume) analyses. An institutional language or ‘speech genre’ (Bakhtin 1986b) is itself a dimension of how a given institutional language is renewed and adapted as it is entered into and coordinates the subjectivities of people at work in particular local settings. On the one hand. and so on – into the actual sites of people’s living where we have to find them as local and temporally situated activities. are doing elsewhere and at different times. namely the social relations that are implicit in its organization. if ethnographers are to properly describe a people’s ways of living. has a long history. writing about how people live. The double dialogue of sociological inquiry Ethnography. originating in descriptions of how ‘others’. Institutional ethnography proposes to address this as its problematic. the varieties of text-mediated discourse. beliefs. and who are marginalized in some way in the society. The project calls on us as sociologists to discover just how the everyday/everynight worlds we participate in are being put together in people’s local activities. It has been deeply embedded in imperialism. . It draws on people’s own good knowledge of their everyday/everynight worlds and does not substitute the expert’s ‘reality’ for what people know in the doing. people not like ourselves. they have to understand the people. including. Institutional ethnography’s radical move as a sociology is that of pulling the organization of the trans. It conceives of the social as actually happen- ing among people who are situated in particular places at particular times and not as ‘meaning’ or ‘norms’. It takes up a stance in people’s experience in the local sites of their bodily being and seeks to discover what can’t be grasped from within that experience. our own. Ethnographers’ descriptions represent them for others ‘objectively’ and ‘as they really are’. INSTITUTIONAL ETHNOGRAPHY 19 organized by and coordinated with what people. Institutional ethnography refuses to accept the terms of such genres as constitutive of the objects of its exploration. The aim is to create a sociology for rather than of people. Rather. of course. the professions. including the concepts. it locates the object or objects of its exploration in the actualities of the work/activity as it is coordinated. the state. on the other ethnographers are committed to betraying these confidences to out- siders who may make of what is told whatever they want. In sociology today it is largely used to describe how others live who differ from ‘us’. theories and so forth that are implicated in that coordination. mostly unknown and never to be known by us. must become to some degree close to and be trusted by them. sociological readers. theory. ideology – the forms of thought in general – are integral to these forms of social organization and relations and are understood as criti- cal to their local replication. as far as is practicable within a given scope of investi- gation. live.or extra-local RULING RELATIONS (Smith 1999) – bureaucracy. Concepts.

The project is fundamentally contradictory. Disciplinary concepts and theories function to regulate sociological discourse and to guard it against this essential risk (Smith 1999). Abu-Lughod 1998). rules of evidence. But as we know now. direct or indirect. The describer is supposed to vanish in the act of writing so that somehow the original of what has been written will appear directly to the reader through the text. Or rather two intersecting dia- logues. methodologies. changed. Classically. from the many critics of anthropological ethnography. this is not possi- ble (see e. socio- logy has sought devices that would enable its accounts of the social to pre- tend to stand outside it. It can never insulate itself from the dynamic of an object that refuses to remain an object. In dialogue with others we are captured. always imperfectly. The issue of power lies in this intersection. by writing over or reinterpreting the various perspectives. one with those who are members of the settings to be described and the other with the discourse our description is to be read in. of some kind or another with others in the world they share with them. It is in the nature of their subject matter that they are exposed to capture by perspectives and ways of thinking other than the sociological. sociology. Dialogue number two is the dialogue between ethno- graphers and their readers. come to see things differ- ently. Indeed it looks as if the ethnographic act aimed at describing people’s ways of living is an oxy- moron. despite its claims to objectiv- ity. Out of the primary dialogue with people who constitute both the resources for the accounts to be written and their ultimate users. can never achieve it.20 PUTTING THE PRACTICE INTO THEORY Description commits an invisible mediation. constrained by its conven- tions. the people they write for. discursive objects and other aspects of the ‘order of discourse’ (Foucault 1981). That discourse has already shaped the dialogue with the people whose lives they are describing in the choices of topics for their interviews or what they are attentive to in their observations. Clifford and Marcus 1986. of the primary dialogue into a single . However. we fashion a secondary dia- logue within the order of sociological discourse. experiences. Part of the problem I’ve described comes from the difficulties created by working up what is essentially dialogic into a monologic form (Bakhtin 1981). that is. Dialogue num- ber one is the ethnographer’s interviews with informants or observations of people’s everyday lives (observation is no less dialogic though the ethno- grapher doesn’t speak). ways of using language. Ethnographers write about their research in dialogue with the discourse in which their study origi- nates. Nor can sociologists simply segregate the sociological from the other-than-sociological dialogues they carry on. Ethnographic work is explicitly a dialogue. In the view put forward here. it confronts a fundamental difficulty in sustaining this: sociological inquiry is necessarily engaged in a dialogic rela- tionship with those it studies.g. The ethnographer’s power is to take what people have to say and to reassemble it to appear in quite a different setting in a different lan- guage and with interests and purposes that are not theirs. Sociology is peculiar in that it aims at understanding the same world that sociologists are part of and do their sociological work in. Sociologists are in dialogue.

beliefs. The insti- tutional ethnographic approach to the necessarily dialogic of any ethno- graphy is one that recognizes and works with it. write and so on) in par- ticular local settings at particular times. This ontology of the social includes in a single ontological realm the stan- dard dichotomy that lays practices on the one side and forms of conscious- ness. They are always somewhere at some time. They occur in time and are done in the particular local settings of people’s bodily being. INSTITUTIONAL ETHNOGRAPHY 21 overriding version in which the differences. on the one hand. appear only as expressions or instances of the dominant discourse. of course. say. and so on. institutional ethnography moves away. In emphasizing the concert- ing of people’s activities as its focus. They are the expert practitioners of their everyday worlds. but the experience of separation from local actualities is itself produced right there in them as people adopt a disciplin- ing of the body so familiar we pay no attention to it and as they take for granted the text as their medium of access to the beyond-the-local. The aim of the enterprise is to be able to return to those who are situated as were the interlocutors in the same institutionally ordered relationships. theories. on the other. from concepts such as Pierre Bourdieu’s concept of habitus (Bourdieu 1990) that reduce the social to properties of individuals or con- cepts such as social structure or system that reify the coordinative dimension of people’s activities. Its aim is not to describe how people live or the meanings they share (Emerson et al. The project is analogous to cartography. with something like a map of how the local settings of their work are organized into the relations that rule them. the object of institutional ethnography’s attention. neither exclusively on the activities nor detaching the coordinating as ‘system’ or ‘structure’. 1995). Language. Institutional ethnography’s people are always embodied. concepts. thinking are here all recognized as among people’s activities. is conceived of as arising in people’s activities (what they do. if they are registered at all. those who directly participated. they know how they go about doing things. ideology. Concepts . concepts. The social is a focus on what is actually happening. The institutional ethnographer’s interest is in learn- ing from them first and then beginning to locate in their accounts the junc- tures between the everyday worlds as they told them and how they are hooked into relations that connect them beyond scope of experience. The social is a focus on activities as they are coordinated. Finding the social The social. It should produce accounts of the social relations and organization in which the doings of the people talked to are embedded that will enlarge individuals’ perspectives beyond what they can learn directly from their participation in the everyday/everynight world. it is to be discovered in people’s doings in the actual local settings of their lives. It receives people’s accounts of their everyday life experiences as they tell them. Thought and mind may be experienced as divorced from the local and from individuals’ bodily being. including.

perhaps better. differences of perspective. Since coordinating and concerting are the stuff of the social. Once we are free from the constraints of belonging to and subordinating our investigation to the dictates of one of its ‘schools’ and governing what can be found by its conventions.22 PUTTING THE PRACTICE INTO THEORY and theories appear extra-temporal on the page but in actuality they are people’s doings in their reading and thinking and in their talk in particular local settings and at particular times. That remains to be discovered. Sociology provides grand resources to support the elaboration of any preliminary formulation as inquiry goes forward. Institutions. as the ongoing coordinating or concerting of people’s activities is a minimal theoretical move. facts. The language. computerized or otherwise replicated. Social relations and organization generate difference. It makes no commitment to what may be found. George Herbert Mead’s (1947) work provides an interactional theory of symbolic behaviour (language). based in her or his body which situates consciousness in a site no one shares. the object of our investigations. Hence the institutional ethnography relies on the language in which people speak of what they know how to do. and so on are expected. and conceive things differently. administrative and legal rules. Institutional ethnography begins with and takes for granted that people experience. Divergence is primary: consensus is a chimera. Nor does it make an a priori commitment to a particular level of abstraction. It locates only a point of entry. printed. Conceiving of the social. And again. there- fore. the speech genre of the institutional setting carries institutional organization. The architecture of . All the phenomena in language. and of how they get things done. see. information. The theory of Marx’s major study of Capital brings into view and analyses the peculiar properties and dynamics of the social relations among people arising in the exchange of money and commodities. interest. we can draw on what comes to hand in our cartography. formulates and makes visible the concerting of people’s talk in conversation. are included in the institutional ethnographer’s object of study and they are indeed of special importance as coordinators or organizers of people’s divergent consciousnesses. Specialized theories recognize and analyse differ- ent levels or aspects of the social. In institutional settings and hence of special interest to their ethnographer are those socially organized forms that generalize and objectify since these must subdue and displace the particularity of individ- ual perspective that arises spontaneously in actual work settings. as objectifying forms of concerting people’s activities are distinc- tive in that they construct forms of consciousness – knowledge. The articulations of the social organize multiple layers of diverging locations that are mediated to people through their activities and people’s activities themselves organize perspec- tives that diverge in the very process of their concerting. Foundational to these forms of consciousness are TEXTS. Each individual begins from the null point of consciousness (Schutz 1962). at a level prior to conversational analysis. Indeed coming at things dif- ferently is what makes the concerting of people’s activities endlessly open- ended and productive. for example. and so on and so on – that override individuals’ perspectives. Ethnomethodology’s conversational analy- sis. Or has been discovered. of their experi- ence.

The first of these is research which Alison Griffith. or pictured of it. It means not knowing what part to play in the dialogue. In what follows. shapes it. These guide the direction of inquiry. The problem is only overcome by somehow finding a direction (and some despairing graduate students. But sim- ply observing is disorganizing. It describes part of a study by Ellen Pence in which she introduces texts into an investigation of the work organi- zation of the judicial processes focused on cases of domestic abuse. specifically provides for how people as individuals become both individuated and invisible in the interchange between money and commodities). concerns or problems that are real for people. In a sense the collection of data and their analysis aim at discovering just how the institutional is being produced by people at work in the particularities of their everyday/everynight lives. Objectification and generalization are themselves the local practices of people’s everyday/everynight lives and are to be explored as locally achieved (Marx’s conceptualization of the economic forms of social relations. Any kind of storytelling selects from the raw material. it is always over. It begins with some issues. written. what to ask. and so on. The storytelling is achieved in actuality whether in writing or reading. I have adopted the device of describing two studies in some detail. I was sent out simply to observe. DOING INSTITUTIONAL ETHNOGRAPHY Finding a direction Of all the possibilities of ethnographic focus. The second study to be described focuses more on the textual dimensions of institutional processes. never do). It is always more and other than anything that can be said. INSTITUTIONAL ETHNOGRAPHY 23 institutions is through and through textual whether in print or computerized and institutional ethnography increasingly incorporates attention to texts and textuality. it is always other. and creates discursive order. what to select. what is to be systematically recognized as relevant to the institutional ethnographic project? The lived world can never be exhaustively described or enumerated. Ann Manicom and I were doing some while back exploring the relationship of the work that mothers do in the home to the educational work of the school. Institutional ethnography begins by locating a standpoint in an institutional order that provides the guiding perspective from which it will be explored. The observation and interviewing that go into ethnographic study need a general formulation of what we would be attending to and recognizing and in how we analyse them and the relevant institutional texts. In that study we relied almost entirely on open-ended interviewing as a research method. but the actuality always escapes. Alison Griffith and I began with our experience . for example. Observation without an attentional frame is anomie. It means not knowing how to look. working with this recipe. what to follow up. When I went out as an innocent graduate student to do an ethnographic study of a state mental hospital in California.

On these walks. Over the period of two or three years before we decided to undertake the research. It is a complex of relations rather than a definite unitary form such as a corporation. of pushing teachers too hard. From experience to institution We decided to work with a procedure which would first establish a particu- lar standpoint located in women’s everyday lives in the institutional context and hence in the ways they are related to and participate in it. On long walks through the RAVINES IN TORONTO we shared the stories of our mothering work. Our explorations opened up the social relations and organization of schooling as those in relation to which women’s work as mothers is done. complaints. we shared confi- dences. Alison had already laid the groundwork for our more system- atic and sociological reflections on our experience by her research into the ideology of the single parent family and its various uses by educational psy- chologists. of not pushing them hard enough. educational administrators and teachers (Griffith 1984). of our children’s struggles. An institutional order doesn’t offer a ‘natural’ focus. in addition to whatever political interests the sociologist may have.24 PUTTING THE PRACTICE INTO THEORY as single parents. Their interviews would give us a standpoint from which we could open up exploration from the side of the school and school board. Because our own experience had been as members of a ‘deviant’ category of families vis-à-vis the school system. miseries and guilt arising from our relationship to our children’s schools. We had learned over the years how to interpret ourselves as imperfect families and we knew we were seen just so by teach- ers and administrators in the schools our children attended. We decided we wanted to understand why we had the kinds of problems with the schools that we identified with being seen as defective parents of actually or potentially defective children. we also framed our collaborative research project on mothering for schooling. We knew that problems our children might have in school would be and were interpreted in terms of their defective families and our experience had been that being labelled in this fashion undermined our ability to be effective in dealing with their schools. we were interested in learning more about the ‘normal’ family against which our own had been seen as wanting. Now we thought we could find out more about what was so special about the ‘Standard North American’ family (Smith 1999) of father earning the income to support a wife and mother at home and children in school if we talked to mothers with children in elementary school about the work they were doing in relation to their children’s schooling. We would start by interviewing intensively a small number of women with children in elementary school. locating a specific institutional standpoint organizes the direction of the sociological gaze and provides a framework of relevance. The interviews in explicating women’s EXPERIENCE of . The investigation itself builds from one stage of research to the next on the basis of interviews. Hence. of our fears about interfering.

Why six in each? We certainly did not claim to have drawn a sample of any kind.e.1 Dialogue in practice Institutional ethnography isn’t about explaining people’s behaviour or about testing theory-derived hypotheses by relating variables derived from individuals’ responses to structured questions. Our interest was not in the effects of different mothering practices on the achievement of the individual child. For these reasons we decided to talk to women whose children were attending schools in middle-class and working-class school districts (Walker 1990b). What defined the textual community thus to be created was the institutional regime we and the women we would talk to participated in and were defined by. We could not know the specific character of the problematic of the everyday of mothers’ work until we had explored it with them. We were interested in how that might translate into differences in the organization of schooling. We began with mothers and the work of mothers in relation to their children’s schooling. We did not think we could learn . The problem is not to select a sample that can properly be treated as representing a particular popula- tion. six with children in an elemen- tary school in a district largely populated with people in a low income group and six with children in a school in a largely professional-managerial popu- lation. We had specifically sought to locate ‘non-defective’ families. but in different forms of the social relations set up between varying indivi- duals at work under varying conditions and the characteristic standardizing of the institutional regime. would specify the problematic of the research beyond its tentative formulation in our own experience. so would schooling. That done. If mothering differed. Alison and I had a specific agenda. The move- ment of research goes from an exploration of the everyday particularities of women’s work as mothers to exploring the generalizing and generalized relations in which each individual’s everyday world is embedded. Institutions are themselves generalizers and their ethnography looks for the ways in which the particularities of people’s everyday doings bring into being the distinc- tively generalized forms of the institutional order. families in which both parents were present. i. Locating people to talk to involves making decisions about the perspective from which an institution is to be explored. Alison’s and mine. nor to be able to make statements of relations among variables that can be generalized from a sample to the population it represents. we could go on to open up the institutional order to which their work contributes and with which it is coordinated. INSTITUTIONAL ETHNOGRAPHY 25 their work as mothers in relation to their children’s schooling. of the same kind of work. We thought that women’s mothering work would vary not only from indi- vidual to individual. the second stage of research would situate their experi- ences in the organization of the schools attended by the children. but also according to different conditions of mothering. After we’d talked to the women. We talked to twelve women in all.

But in principle institutional ethnography might begin with one individual. Meaning and subjectiv- ity are not differentiated from accounts of people’s doings. For the very open interviewing we had used at first. we wanted to see how it was done where economic conditions made it more difficult. Interviewing One of the problems of sociological research in institutional settings is that interview respondents speak from the generalized and generalizing dis- course. It focuses on the concrete and everyday and particularizes rather than gen- eralizes. has the major merit also of evading the divorce of subjective and objective that often requires the sociologist to hover unhappily between objectified descrip- tion (as in ethnomethodology’s conversational analysis) or concepts of mean- ing which are generated by methodological apotheosis (as with grounded theory). Hence in devising our new interview schedules Alison and I had to find a format that would anchor the interview at the descriptive level and avoid elevation into the institutional (Smith 1990a). When our initial research plans and progress were disrupted. importantly. No doubt there would be more systematic procedures than those we used in this early study. Alison and I had to plan our next steps with greater economies of time in mind. Our interviews in . Choosing a standpoint from which the insti- tution will be explored is a key step in the ethnography (Smith 1987: 181–190). We did not only want to explore that work in settings where it was economically supported.26 PUTTING THE PRACTICE INTO THEORY enough without being confronted with differences in the conditions under which women’s work as mothers was done and hence with having to recog- nize the ways in which the generalized/generalizing relations worked on and organized how women managed the ‘schooling’ dimensions of their work under varying conditions of that work. we substituted a more formal- ized traditional approach using more or less set questions. Others may be interviewed but those interviews will not establish a new standpoint though they will. Such discourses are extraordinarily lacking descriptive content and can be largely useless to interviewers unless they are investigating the order of the institutional discourse itself. in our sense. Even when they talk about their own work. Marjorie DeVault (1991) in her study of feeding the family demon- strates how interviewing people about their work in the generous sense of the term used here elicits talk about how they think and feel about the work of feeding the family as well as about the practicalities. they may move into their everyday competence in the ideological language of the institutional dis- course. ‘Work’. Institutional ethnography is sampling an institutional process rather than a population. The presence of institutional organization is in what someone has to say about her or his work (in the sense that we use the term here – see Glossary). provide a different pers- pective on the major themes that have emerged from interviews with those who make up what we might call the ‘standpoint’ sample. Here is where our concept of WORK kicks in.

to analyse it as a ‘social relation’ (discussed at greater length in the following section). asking them for each step about the work they did in relation to it: what was involved for them in getting their child off to school? Did they come home for lunch and what had to be done (prepare meal. if indeed they are allowed to watch first thing. Or rather it expressly recognizes a dialogic of social inquiry that is always ineluctably there. The formalized structuring of questions used in survey research and in some forms of qualitative research (though now by no means all) suppresses the effect of dialogue. and so on. But we also talked to women for whom achieving the regular- ity of the school day was not salient. although we were aware of the problem of the assumed typicality of the day. we had dis- counted that in favour of the access to the concreteness of the everyday work of seeing to it that children have eaten breakfast. even when the same topics or questions are introduced each time. In our dialogue with our respondents. We had not seen. and so on)? When the child/ren came home from school. the school day’s orderliness as an intersecting of the work of school teachers and administrators (as well as. get child back to school on time. We began in this way to see the typical schedule of the school day as a joint product and. aiming at the outset to produce a monologic (Bakhtin 1981) . are dressed appropriately for school and weather. There is progression from interview to interview. from each interview what may inform and change the subse- quent (DeVault and McCoy 2000). the law) and of mothers who sustained its scheduled character by the work they aimed at achieving it. INSTITUTIONAL ETHNOGRAPHY 27 the new setting opened by taking respondents through a standard school day. our thinking was changed and some- times in ways that were only contingently related to the planned relevances that guided interview topics. the researcher is permitted to learn. that they are pried away from television. what then? And so on. Research as discovery In institutional ethnography. as we came to see later. our interviews did not seek to make up a collection of data we could wrap up in a box of code. It pushed us to think further than we had at the outset. and. though it now seems obvious. And to those who expressed the strain of coordinating their own and their partner’s schedule of paid work to the unaccommodating school schedulers of their children. Since we were not following the traditional sociologi- cal procedure of starting from disciplinary concepts or theory to specify first questions and then coding categories. ultimately. We found what we had not known how to look for. One such moment led us to become aware of a dimension of our work and thinking that we had not recognized at the inception of our study. The concept of SOCIAL RELATIONS proved more analytically powerful than we had anticipated in those days when institutional ethno- graphy was still relatively undeveloped. perhaps must learn. We’d thought of the orderly para- meters of the school day simply as a data-eliciting device. eventu- ally.

But that did not excuse us. In the course of inter- viewing. we encountered situations that made us. Normally such responses are ignored and certainly not considered relevant to the investigation. namely that the parties to it take the risk of being changed. For the most part sociological method expressly suppresses the dialogue at work in what we call our ‘data’. if you like. But institutional ethnography. even anxious. Such amplitude of time let alone of other resources precipitated feelings of guilt. including taking them to see Shakespeare plays at the Stratford (Ontario) Festival. another moment of change was particularly significant. Furthermore.28 PUTTING THE PRACTICE INTO THEORY in which the respondent’s part is subdued to the terms of the pre-set questions and the pre-coded responses. and had much time to spend with her children. not only for the specifics of our study. Do respondents in an interview risk being changed by the interview? In hearing themselves focus in their talk on what they have not perhaps imagined would be of interest to anyone. Tracing this research byway. what the researcher must learn from him or her. as single parents. did not work outside the home. But institu- tional ethnographers are actively seeking to be changed. who had not had time or other resources to do this kind of work with their children. anxiety and pain in the researchers. can be successful only if it takes advan- tage of what Hans-Georg Gadamer (1994) takes to be essential to dialogue. in these terms. She told us that her children helped choose which plays they should go to. We were respon- sible whether the conditions of doing that work were present or not. rather than leaving things at the stage of telling and sympathetic listening. we came upon what we now call the mothering discourse (Griffith and Smith 1987). we started to question why we reacted in this way to interview experiences of this kind. From historians we learned that this . Recognizing interviewing or observa- tion as essentially dialogic recognizes the researcher’s interests in the research as integral to the dialogue while at the same time relying on the other to teach. uncom- fortable. how to think differently about what they are learning. they are doing discursive work and hence their change aims at change in institutional ethnographic discourse and more generally in what we know about how institutions work. Sociological research. One such was that of an interview with a woman who was the wife of a professional. at that time 7 and 8. as they go about their work. So. to discover not only what they did not know but also. are they changed? Is it something that can be made explicit? I doubt it. Dialogue is con- cealed either by the use of data that has suppressed dialogue before it arrives at the analytic site or by deploying theory that converts the many-voiced into the monologic (Bakhtin’s term). These are the normal risks of conversation. For Alison and I. can take advantage of such responses as opportunities for opening up dimensions of the institutional regime that weren’t recognized at the outset of the project. She described her many activities with her children. but also because it introduced to us the idea that institutional ethnography did not ‘naturally’ confine the researcher to the planned parameters of the project. recognizing that researchers are in the same world as that they are investigating.

either practical or emotional or of knowledge. One interview I did myself returned me to a familiar experience. women in the middle classes were mobilized by the mothering discourse to commit themselves. socialization and education. Now in my interview with this particular respon- dent. I felt at each question that I was somehow interrupting her train of thought and that my questions had no continuity with the preceding responses.’ These words were echoed in countless publications through- out the period. I was having the same experience. I had the uncomfort- able sense that I could not line up my questions with where she was coming from. Some while ear- lier I’d interviewed my mother. This is her responsibility and her opportunity. Mothering discourse places on mothers an undefined responsi- bility for their child’s educational achievements. We discovered also. under the guidance of profes- sional experts. on what she should have contributed to the child. in the early twentieth century. Even though I paid due respect to what she had to say and listened carefully. I came to have an understanding of that period that I’d not had at the time of our interview. At a later time. A 1936 article in Chatelaine [a Canadian women’s magazine] warned mothers that ‘your child mirrors you and your home: if your child is a problem child. Any problem that the child might encounter is read back into what she has done or failed to do. I was not able to locate myself pro- perly in the universe of discourse from which she spoke. Looking back I can .’ (Arnup 1994: 151) The mothering discourse has changed over the years in its specifics. to a demanding work of child rearing. I knew then that it was my ignorance that structured the interview. then in her mid-eighties. though few. about her experi- ences as a young woman in the women’s suffrage movement in England. credentials and career-structured occupations and the profes- sions. It was associated with the rise of the new middle class organized around education. their responsibility knows no limit. who clearly were not participants. In a sense. personality and general well-being. somewhat to our chagrin. particularly in relation to their sons). INSTITUTIONAL ETHNOGRAPHY 29 discourse emerged. in North America at least. that the conceptions built into our interview strategies presupposed the mothering discourse and that many of the women we talked to were also participants. Alan Brown reminded readers that ‘the mental environment of the child is created by the mother. There were also those. Although I was a reasonably experienced interviewer. after she’d died and after I’d had an opportunity of teaching an undergraduate course that I’d had to prepare for by reading extensively about the women’s suffrage movement in Canada and the United Kingdom. For participants in the mothering discourse. probably you are a problem mother. but not in the onus of responsibility it lays on mothers (today we begin to see fathers implicated too. They knew how to ‘talk’ the discourse just as we did and we have no idea how deeply our shared discursive competence went into shaping the stories we were told. A mother’s responsibility is not determined by specific tasks but by outcomes. In the 1932 edition of The Normal Child.

or how to use pencils and crayons. the less time to teach the curriculum. the more time that had to be given to teaching ‘background’ skills. Her work emphasizes the implications not for the individual child but for how teaching can proceed in the elementary classroom and hence the implications for the level at which the curriculum can be taught. Non-participating mothers did not have the same commitment to the work that would complement the teachers’ work in the classroom. Institutional ethnography is. In a school in a low-income district. On the one hand there were mothers at work in their homes under whatever conditions of time and other resources available to do the supplementary work of teaching and sustaining the orderly process of the school day. the teachers she spoke to reported that they had to put considerable classroom time into teaching children such skills as how to turn the pages of a book in the right direction and the direction in which the writing is read. some had not acquired the background competences that could be taken for granted by teachers in the school in the middle-class district.30 PUTTING THE PRACTICE INTO THEORY see the effects of the mothering discourse in how I assumed that all mothers would be primarily oriented to a child’s successful school career and hence did not attend to the possibility that for some such issues as the economic well-being of the family or a sense of the importance of local and family rela- tionships might be given priority (these are my guesses after studying the transcribed interview many times). Ann Manicom’s (1988) study of the experience of elementary school teachers in the classroom opens such a window in its exploration of teachers’ work in schools in differing class contexts. I failed to engage with her in the kind of talk that might have opened up other ways of thinking in me. Exploring an institutional regime is best done from more than one perspective. Analysis does not work with a set of categories and concepts that are laid over the texts of our transcribed interviews but locates what we have . Quite simply. in principle. never completed in a single study. Working with the data In institutional ethnography. The mothering discourse allowed no such alternative priorities and I did not attend to that possibility. The mothering dis- course earlier described was clearly of importance in coordinating the efforts of participating mothers with those of teachers and of the school in general. each giving a different view of the terrain. and on the other. the concept of social relation is central to the analysis. Relating her study to ours brought the social relations of mothering and schooling more clearly into focus. In the absence of my awareness of the problematic status of this assumption. Inquiry is conceived overall as opening different windows based in how people are positioned in the institutional regime. there were teachers whose conditions of work in the classroom were shaped by the time and whatever other resources were available in the home. While the children they taught might be skilled in ‘street smarts’.

Manicom’s (1988) study has given us a stronger sense of just how important this work is for the school in general. The daily scheduling of the school’s work by teachers. As George Smith (1995) emphasizes. It is also available. perhaps a schoolmate’s teasing. however. we look for what people say about the ‘work’ they do that connects them to the work others are doing elsewhere or elsewhen. but one that examines how work in one institutional site connects with the work of others in other such sites. or the like. known and unknown. Schools rely on the extensive supplementary educational work that many mothers do at home. reading. familiarizing the child with letters using the plastic alphabet with fitted with magnets that can be stuck to the refrigerator. we believe. This is the work that the mothering discourse calls women to. Then there are the less visible forms of work. It may involve such familiar tasks as helping the child with homework. known and unknown. The interviewing procedure has this analytic strategy built into it. INSTITUTIONAL ETHNOGRAPHY 31 learned from respondents about their ‘work’ as it is coordinated by relations connecting it with the work of others. Such mothering work is mobi- lized by the mothering discourse. such as that of monitoring the child’s school experience by asking her or him about what happened at school that day or scrutinizing work brought home. It also includes less formal aspects like teaching the child to iden- tify colours. sometimes both parents. In the interview transcripts. find in their talk their particular moments of participation in social relations that hook their local experience to the work of others elsewhere. tutoring the child. visits to libraries. the concept of social relation does not identify a class of phenomena. ‘speech genres’ to use Bakhtin’s (1986b) term (LANGUAGE AND SOCIAL ORGANIZATION). let alone for the individual child’s career at school. Mothers. We can. Hence social relations are already implicit in how people talk about their work. It can use up all the time that is available – and still leave a mother feeling that she is wanting. museums and zoos. We take the view that social organization is built into people’s ways of speaking or writing. therefore. Mothers must also sometimes spend time dealing with a child’s anxieties about school – perhaps a teacher’s speaking to her or him sharply. but mothers in particular go on school trips and volunteer time to the school. The school day is thus seen relationally. usually done when a child returns home from school in the after- noon. but focuses on how what people are doing and experi- encing in a given local site is hooked into sequences of action implicating and coordinating multiple local sites where others are active (Smith 1990b: 93–96). The generalization of the school system has come to be complemented and sustained by a generalized discourse that assigns to mothers this kind of supplementary educational work. in particular interviews in the ways in which people in how they speak incorporate the socially organized rela- tions in which their experience arises. school secretaries and school administrators is . Hence analysis isn’t a mock-up of a statistical procedure. it is a service to child and school that lacks specification and boundaries.

i. according to the school board and school administrators we spoke to necessarily typical of the downtown community where there was a relatively high proportion of single parents. these were families which were ‘intact’. women.32 PUTTING THE PRACTICE INTO THEORY complemented by the work of family members. they are generalized across individual situations and experiences. all the women. For the most part. in which both parents were present and the latter were not. It is hard for them to make the objective of a child’s success in school the priority around which the household day and resources is managed. She was the only single parent in the group of women we talked to. Single parents. There was one in which the family organization was just the same as most of those with children attending the middle-class school in that the husband was employed full time (he was an engineer) during the daytime and the wife was not in paid employment. When the school day is examined relationally. One. are ‘institu- tional’. Institutional ethnography recognizes that the institutional forms . the school day. Appar- ently the school was not aware of this in giving us her name as someone willing to be approached for an interview. in various ways were engaged in earning a living. There are the schedules of paid employment that don’t coordinate with the schedule of the school. But in all the others. nor in general for the supplementary educational work on which the middle-class school relies. Families with the children in the school in the lower-income area had rather different patterns of organiza- tion. that is. their husbands were in paid employ- ment outside the home and they were not. no parent was at home full time during the day. A couple worked a few hours a week. from the side of the school. Though individuals and how they go about their work is indi- vidually various and unique to each. So the time of the women with children in the lower-income district school was not available in the same way at the times that were fitted to the school day. we can also see how the realization of the province’s school curriculum as an allocation of daily time- slots depends on the family’s commitment to the work of ‘getting to school on time’. with one exception (she was in full-time training as a veterinary assistant). We are talking about the work and experience of individuals. But back in the home is another hook-up. comes into being as a relation that is re-created everyday on one side at home and on the other at school. What seems to have this firm and finalized existence. are likely to be less well situated in terms of time and resources than the mothering discourse model requires. as was our own experience. but the rela- tions in which they are active and to which they contribute. they are active in producing together objectified forms that transcend individuals and are both historical and allochronic. though not all are in regular jobs. The school day. is the framework within which curriculum space is allocated to differ- ent parts of the curriculum. children and sometimes men. Here was a dif- ference between the women whose children attended the school in the middle-class district. Of course. The school day presupposes a parent available at home to manage the sched- ule of the ‘home day’ so as to coordinate the uncoordinated.e. did work full time. who actually turned out to be a single parent.

class size and/or the teacher–student ratio increases and teachers have less time to spend with individual children and. All responded. parent and teacher to the societal level of the social relations of class. though in a new way. It is also possible to locate specific possibilities of change (Pence 1996). An important dimension of the avail- ability of teachers’ educational time is class size. how time is allocated is central to how a public educational system produces inequality. including the functional foci of specific insti- tutions. Imagine a total educational work time. Yet they were all in various ways active in producing and reproducing the institutional character of the schools their children attended. exploring the institutional order. As less public money is spent in education. we’ve found that tracking the interchanges of the-time-it-takes is analytically valuable in sorting out the interconnections (we note though we have not explored the theoretical and analytic linkages to Marx’s conception of labour time). almost with the same words: ‘Oh. We can begin to see then how the public school system comes to operate as an ‘engine of inequality’ (Smith 1997). in shifts in government policy. to be allocated to either school or home. Where time is not available at home to do this work or when that work is not done for whatever reason. or the like. In this way. In education. do not have the time to do ‘repair’ work with a child who may be falling behind her or his classmates in a particular area. of course. in changes in it. and in more general changes in economic organization that are taking place. is grasped from a standpoint within them and is brought into view in how these relations coordinate people’s local practices. We were dis- covering that the public school system functions in North America as an ‘engine of inequality’ and we could begin to describe how people are putting together this engine in the practices of their daily lives as those practices are coordinated with one another to create social relations that the different . each family’s way was an idiosyncratic configuration of particular people. Class size is regulated sometimes directly but always through funding. The women we talked to each had their own way of organizing the work of schooling that is done in the home. institutional ethnography opens up into larger social relations. It is this insti- tutional character of the social relations coordinating the work being done by mothers and by the professionally and administratively regulated work of teachers that enters the unique and particular relationships between child. in particular. We asked the teachers we interviewed how it would be if their class size was halved. as by ask- ing parents to spend time reading with the child. the school does not expand its time to compensate. the ten- dency is for the teacher to pass the problem back into the home. discovering class. In this situation. then we’d really be able to work with the children as individuals. We can begin to locate people’s everyday lives in the institutional order. Current assumptions (built into curriculum expectations) are that the supplemental education work mobilized by the mothering discourse will be done at home. The complex of the ruling relations. INSTITUTIONAL ETHNOGRAPHY 33 and relations are always being produced by individuals at work and yet that they are generalized and standardized.’ We were. In tracing these relations.

In the economic conditions that govern the availability of time. not everybody reads a given text in the same way. again not as a matter of an individual mother of a particu- lar child but as a consequence for the school itself and how it could deliver a standardized curriculum. so that one aspect of the social relation (in the sense used here) in which both parents and teachers are active becomes visible. In an ethnography therefore texts are taken up as they enter into action. we discovered how that work was mobilized and organized by a mothering discourse that claimed women’s time not just as something done for her children but as something done for the school her children attend. Texts as coordinators The grand puzzle of institutions is how they generalize across many local settings of people’s activities. through this work. the printed text. but as ‘organizers’ packaged for transmission to multiple sites.34 PUTTING THE PRACTICE INTO THEORY parties may not be aware of. We have. In a sense. Texts as read and written in the everyday actuality of people’s work coordinate what people are doing in one local setting with work done by others elsewhere or at different times. We had opened up an institutional process to discover a work of women and the conditions of its doing. But making visible a more extended organization of social relations is not the only or indeed the primary object of institutional ethnography. They can be seen in the quotation above from a historian’s account of what Alison and I have called ‘mothering discourse’ (Griffith and Smith 1987). We were replacing this with an examination of the interlocking of work. Texts bring external regulation into the immediacy of the everyday/everynight world. we could see home and women’s work as a kind of conduit that transmitted those condi- tions to the school. Of course. However they may be read. time and economy of actual people situated differ- ently in an institutional order. we were unpicking class itself against those theory-derived conceptualizations that create virtual collections of people in a virtual structure. The social act unfolds in time as well as in actual local settings. and do not intend. Institutional ethnography argues that general- ization is text mediated and that standardized and replicable texts coordi- nate the local settings of people’s work. the computer software or other textual form bring an identical set of words or images into local sites. The tricky matter for the ethno- grapher is to locate texts in time and as constituents of social relations. The texts that . been able to talk with women about how their work as mothers is located and show them something of the relations they are caught up and are active in. Our analysis shifts responsibility from women as mothers not to the teacher but to how teachers’ time is allocated and to the issues of funding and class size that are consequential for the conditions of teachers’ work and ability in practice to treat each child as an individual. Their conceptual dimensions are held not as meaning. but for every site into which a given text is inserted one side of the text–reader conversa- tion is fixed and unchanging from site to site to site.

Readers in different sites and at different times can engage with and be regulated by the ‘same’ text. The printed or otherwise replicated text stan- dardizes for any reader one term of the text–reader conversation. Each textual step is the basis on which the next step is taken. predetermined and remains unchanged by the history of its reading. One ‘party’ to the conversation is fixed and non- responsive to the other. in a sense becoming its voice. Spoken conversation takes shape as each speaker responds to the other whereas in text–reader conversations. Ellen Pence was part of a group in Duluth. We might have examined them at a number of points at which they coordinate parent–school relations – the report card. the other both takes on the text. social workers and so on. articulating them into extended social relations beyond. even its agent (Smith 1990a. and acts from it. 1999). referenced. In a sense. Rather she examines the interlocking textually coordi- nated sequences of action that relate the law. or in the texts of the mothering dis- course – but we did not. judges. The standpoint she adopted was that of women situated in (and in many ways outside) the judicial process whose partners had been charged with . one side is fixed. Those working in this way had ample experience of how little even the improved efficacy of the judicial process contributed to women’s safety. The work organization is not an effect of a single unit regulating the various functions within a corporate entity. The text–reader conversation is active. must be recognized as ‘occurrences’ at the moment of reading that enter into the reader’s next doings or ‘responses’. Minnesota. we were scarcely aware of the significance of texts in the organization of institutional relations. Or they are at work in talk or in writing/reading as organizers of local settings. responds to. that had been pro- viding advocacy support for women whose partners had been charged with domestic abuse. interprets. probation officers. for example (Stock 2000). the words that organize may thus be spoken in just the same way to different people at different times. the work of the police. Here then I draw on a study by Ellen Pence (1996) of the judicial process in cases of domestic abuse. women’s advocates. At the time that Alison and I did our research. She never sees the problems as those of individuals. Reading is interactive. The text is both activated by the reader and inter- preted in relation to the local work organization and courses of action of the setting in which it is read. making the texts so ubiquitous in the judicial process part of her observational and analytic focus. The work in each site is coordinated with others through the texts that each is legally responsible for producing at various stages of the overall process. aimed at. governing. INSTITUTIONAL ETHNOGRAPHY 35 coordinate local settings and particular times. Pence’s study drew on that experience supplemented by rare oppor- tunities for observation and participation. lawyers. the ongoing development and concerting of activities. Her study of the judicial process traces it from the moment when the dispatcher receives a 911 emergency call to sentencing. but it is also peculiar and unlike conversation in face-to-face talk in that the text remains the same thing no matter how many times it is read. and at the same time.

He shows the police report as constructed to select those aspects of what the police saw and how they attended to what could be fitted to fit the law under which the men were charged. a Friday: A domestic between a co-habiting couple occurs in the city. George Smith (1990a) has made a detailed analysis of this step in his study of charges brought against gay men engaged in sexual activity in a bath- house in Toronto in the 1980s. 4: The report is transcribed and returned to the Patrol division. the victim and witness information and statements. 2: The sqd. Jan. Jan. 8: It is placed in the city attorney basket. an attorney will review it. 25: The file is returned to the DPD (Duluth Police Department) Detective Bureau clerical person with a summons attached. and so on. A sqd. Jan. She quotes a fictional account created by a detective to illustrate the ways in which the processing of a case can be dragged out by the sequence of steps that are involved. Jan. Some time within the next couple of weeks. 27: The clerical types out the summons information and mails the package out. If someone is arraigned for trial. In the instance of domestic abuse. The report standardizes for all involved what is known of the incident for the prosecuting attorney. this report will have played a central part.36 PUTTING THE PRACTICE INTO THEORY domestic abuse. occa- sionally a neighbour. plays a critical role in coordinating the work of others. the process is initiated by someone present at the scene. It is Friday. as Pence (1996) shows. sometimes a child. . Jan. direct a clerical to fill out the neces- sary forms. 11: It is logged into the city attorney’s office and sent to an attorney. sometimes one of the adult participants. The police responding to the 911 emergency call that initiates the process in most cases translate what they discover in the setting and the events that occur there in their presence into a report. 7: The report is signed by the supervisor and taken to the traffic division where it is logged as a warrant request. It also illustrates a sequence of processing interchanges: Jan. including all of the reports. The work of the police in rendering the local particu- larities of events into textual form is critical in the initiation of the judicial sequence of action and their report. Text-processing interchanges are those work sites in the institutional processes into which a given text is entered and from which it is passed on or incorporated into new texts to the next site which may also be a processing interchange. Jan. decide to issue. The victim gets a black eye and bloody nose and calls the police. [officer in charge of squad] dictates a report. Jan. for the lawyer for the defence. 1. 1993. Pence introduces the concept of PROCESSING INTERCHANGES as part of her analysis of how texts coordinate this judicial process. They go back to their business. for social services (particularly if children are present). [police squad] responds and finds the offender gone and is not able to locate him within the 4 hours.

And. (Pence 1996: 67–68) This account makes vivid how the text of the original police report travels. This may take a week. is added to and incorporated into other texts. the whole of this sequence has implicit reference to the legislation that has established domestic abuse within criminal law as well as to other like texts that regulate local practices (G. At the end of the day. Her interest in using this method of inquiry is to disclose how institutional processes are assembled in the everyday work of people implicated in them. March 31: He appears and pleads not guilty. INSTITUTIONAL ETHNOGRAPHY 37 Jan. Jan. Feb. not worrying and arguing about what will or should happen. how the time the textual processing takes jeopardizes the safety of the woman involved. There it is placed into a basket of complaints to be filed when they have time. 23: It is placed into a basket where it sits for the standard two week minimum grace period. having continued to live together. as does the detective who wrote the memorandum.W. March 26: It is received into the warrants office. taking the standpoint of battered women in relation to the institutional process. let’s say it goes within a couple of days. of course. entered into the computer and placed into the basket for service. and how the accumulating texts coordinate the activities of people positioned differently in the institu- tional order. 22. He then carries it into the Clerk of Court’s office. the court file is carried back into the Clerk of Courts offices. March 24: It is placed into the warrants basket to be picked up by the Sheriff’s warrants office. we can see. logged in. From the standpoint of women who are abused. Nichols carries it to the court. share the first day since the assault that they have not thought or argued about it. March 12: It is received in the city atty’s office and sent to the issu- ing attorney. March 27 and 28: The couple spends first weekend since his failure to appear. Then. Pence’s (1996) work is neither simply analytic nor simply critical. the outcomes of the . A jury pretrial [jpt] is set for the first week of May (jpt’s are always the first week of the month and he is now too late for April). swears to it and has it signed by a Judge. March 18: Sgt. Feb. 29: The offender opens his mail and notes that his court date is set for Feb. March 17: It is dictated as a warrant and returned to the DPD.’ It may sit there until a stack ‘worth’ picking up or mailing back over accumulates but to be charitable. 22: He doesn’t appear for court. March 10: It is removed from that basket and placed into the ‘return to city atty [attorney] for warrant basket. March 30: He calls and agrees to come in the next day. 28: The victim and offender. Smith 1995).

Though domestic abuse comes under the criminal code in Minnesota. Focusing on the mediated role of texts displaces focus on individuals and brings into critical consciousness aspects of how their work is coordinated that are not ordinarily examined and are. Extraneous detail not strictly relevant to determining whether there are grounds for a charge. how their work is concerted so that it produces the outcomes it does. And so on. Such a critical investigation can be the basis on which changes can be . however. Not only their work with texts. For example. The reports may be so sparse that they are inadequate to determine whether a charge should be laid. by examining the textual forms of organization. One place to begin. the reports may be deeply flawed by racism. If people at work in the police and judicial process could examine. The textual. The degree of injury is not always established. There are no established instructions for what must be attended to when the police come to a situation of domestic abuse. at the point of interchange between mothers’ work and the work of teachers and school administrators where texts were not. the scene is not treated as a crime scene. such as that the parties have been drinking. is to introduce standardized procedures for writing such reports which are becoming part of police training and are also required by the supervisor who checks the report. Where those involved are native peoples or African American. of considerable importance in how the institutional process works. has special significance because it allows the institutional ethnographer to locate the essential modes in which translocal and allochronic relations are coordinated and through which the institutional property of being generalized across local sites is achieved. In this brief exposition of one part of Pence’s (1996) study. Choice of approach depends on the situation that the researcher has undertaken to explore and on the forms of coordinating people’s activities that actually obtain there. Pence’s study focuses on the significance of texts and the processing of texts in the sequential organizing of people’s work. for the most part. the pro- tocol for making reports is not well established and standardized across the state. Pence (1996) suggests that those involved in institutional processes can themselves conduct a systematic investigation using institutional ethno- graphy as a critical method for examining the actual outcomes of how their work is coordinating. but also how their work in general is coordinated by texts. Alison’s and my study was oriented to a social relational sequence that. maybe they could also find ways of designing it differently.38 PUTTING THE PRACTICE INTO THEORY judicial process are rarely successful in securing their safety and may even place them at greater risk. directly implicated. are included. we can see how the ‘observation’ of how texts coordinate institutional process complements rather than displaces the explication of the interlocking work of people whose local activities are bringing into being the institution’s translocal and allochronic order. But the analysis of textural coordinators does not mean suspend- ing an interest in people’s work in the local settings of their everyday/ everynight living. at the same time. are integral features of institutional organization. then.

the stress. Indeed that is a distinctive effect of the objectified forms of the ruling relations. their own knowl- edge. They go on from this account to locate such experiences in economic policies of Sri Lanka in the changed global economy and the complex of communica- tion and management created by transnational corporations specializing in textiles. INSTITUTIONAL ETHNOGRAPHY 39 designed and introduced that do not require the radical modifications of a whole that are generally beyond people’s power to achieve. interesting perhaps as such and possibly useful to those whose lives are shaped in the ruling relations and would like them to be different. It might seem as though institu- tional ethnography produces a collection of particular studies. Each study. But each study reaches into and investigates relations that are generalized beyond the particular case. judgement and will. The latter locates Sri Lankan textile factories in a worldwide hier- archical division of the labour of fashion production articulated to different markets (Cheng and Gereffi 1994). Each study examines a facet of the great trajectory of ruling that has been developing and being developed both technologically and in scope and effectiveness since the nineteenth century. job insecurity and ‘the growing casualization of work’ as women experience them. Institutional ethnography aims at uncovering. The studies that have been and are being done explore each from a different site the complex of the ruling relations. Currie and Wickramasinghe (1998) describe the long hours. colour and so on that become a motive to buy. Institutional ethno- gaphy’s characteristic strategies reach equally effectively into the ruling rela- tions of capital. the progressive despoiling of people’s local and particularizing control over their everyday lives as the expansion of the ruling relations continually displaces and expropriates their self and mutually generated relations. From there connections can be made with the organization of design and advertising in the fashion industry that estab- lish the perennially changing and market-differentiating norms of style. textually mediated forms of organization that connect people’s everyday/everynight worlds into the contemporary regime of capital accu- mulation. A study by Dawn Currie and Anoja Wickramasinghe (1998) begins with the everyday experience of women garment workers in factories in a Free Trade Zone in Sri Lanka. The concept of institution is not intended to confine investigation to forms of objectified organization associated with the professions. CONCLUSION Increasingly social organization as it evolves among people in direct rela- tionships is displaced or regulated by exogenous systems of rationally designed. therefore. the effects on their health. brings into view an aspect of the changing regime of accumulation that is assem- bling distinctive institutional forms evolved out of and transformative of . Studies such as those described all too briefly above go much beyond what particular studies may have to tell us. from the stand- point of people located in a definite institutional site.

worked with and become available to the next practitioner. The ways in which power is brought into view as a mobilization of people’s coordinated activities also points to ways in which change can be inserted into organization from within. but also reach into teaching. I use it to designate a class of those relations that organize the local translocally. Gee. And we are all insiders in one way or another. responding to the different properties and features of organization that come into view. Different social classes use the same words in different senses and disagree in their interpretation of events and situations’ (Macdonnell 1986: 3). Each new study. by locating people’s site of experience in the social relations of the institutional regime and explicating what we can of the relations that enter into that experience. Hull and Lankshear’s (1996) use is rather like Macdonnell’s but they extend it to a range of social forms including ‘different sorts of street gangs. A general usage of discourse from which mine diverges is Diane Macdonnell’s (1986). 2001). I recommend it as a skill in an analysis and explication of the form of power in which most people in our kind of societies are implicated. although these are by no means neglected. Exploring the ruling relations from different positions discovers them in greater depth and complexity. we can create something like a map of the relations in which people’s own doings and experiences are situated and by which they are shaped. The products of the work of institutional ethnography not only feed into acade- mic settings. Each study therefore brings into view another piece of the complex of objectified relations and organization.40 PUTTING THE PRACTICE INTO THEORY those that have gone before. When I teach institutional ethnography as a research approach. LIST OF CONCEPTS Discourse My usage originates in Michel Foucault’s conception which appeared origi- nally in his Archaeology of Knowledge (1972) and was subsequently developed particularly in his lecture on ‘the order of discourse’ (Foucault 1981). mediated by texts (print and electronic) in which our daily/nightly lives are embedded and by which they are organized. Ellen Pence has developed institutional ethnography as a method that people working in institutional settings can use to explore and redesign the coordinating of their own work (Pence 1996. Rather. Nor does it explicate power as the ineluctable. into the initiation of change and into awareness of the implications of changes that are imposed upon us. Institutional ethnographic research and analysis does not displace or reconstruct the experiences of those implicated in an institutional regime. has something new to say about how they are put together. elementary schools and . New kinds of analyses are sug- gested. She describes discourse as ‘The kind of speech proper to the shop- floor of a factory conflicts with that of the boardroom.

1996: 10). executives.. Ethnography does not here mean. Their ‘practices’ are coordinated through discourse as how what they do is made accountable to themselves and others (Garfinkel 1967). and analysis of such a complex of relations.. social classes and sub-classes and so on and so forth’ (Gee et al. birdwatchers. I take it up from women’s standpoint as I’ve defined it. or other. [It means] rather a commitment to an investigation and explication of how ‘it’ actu- ally . INSTITUTIONAL ETHNOGRAPHY 41 classrooms. Its methods. Hence the subject is always embodied and is located in an actual situation of activity which the discourse coordinates with others also at work. recollection of work experience. the order of discourse is seen as bringing into being an objectified organization of social relations that exist only in people’s activities but that come to stand over against them. not conceived in the abstract but from the entry point of some particular person or persons whose everyday world of working is organized thereby. The objectification and trans- local organization of discursive relations must themselves somehow be pro- duced and reproduced by particular people located just where they are. My own conception of discourse rejects this generalization of the concept to what used to be described in sociology as ‘subcultures’. are constrained by the practicalities of investigation of social relations as actual practices. I want to preserve Foucault’s (1981) conception of the order of discourse but to stretch it in ways that escape Foucault’s paradigm. And they point out that ‘We are all capable of being different kinds of people in different Discourses’ (1996: 10) (the parallels between this concept of discourse and the ordinary sociological conceptions of role and social system are striking). Since discourse is viewed as organized social relations among people. restriction to methods of observation and interviewing. but also the practices are of definite individuals located just as they are in the actualities of their everyday/everynight worlds as they/we experience them. Not only is discourse taken to be practices (Foucault). as it sometimes does in sociology. academic disciplines and their sub-specialities. interviewing. of actual practices and relations. ethnic groups. descrip- tion. use of archives. overpowering their lives. police. A rule of institutional ethnography’s discourse is that inquiry may not escape people’s essential embodiment and hence their essentially localized (time and place) mode of being in the world. feminists.. Institutional ethnography explores the social relations individuals bring into being in and through their actual practices. This indeed is a central difference between this and other concepts of discourse. genders. I quote from earlier work: The notion of ethnography is introduced to commit us to an exploration. whether of observation. Note however that the institutional ethnography as a way of investigating the problematic of the . . textual analysis.. works. This is the specific relevance of the concept to institutions when the generalization of the work that produces the insti- tution on an everyday basis is problematized rather than taken for granted. Ethnography As in ‘institutional ethnography’.

it depends on this utterly. Rather therefore than the deterministic formulation of feminist post- modernism. To speak from experience was to struggle with discourse and to appropriate terms or force them to behave as they had not behaved before. Everyday/everynight worlds I use this expression as a reminder that women’s work in the home isn’t just a daytime affair.42 PUTTING THE PRACTICE INTO THEORY everyday world does not involve substituting the [sociologist’s] analysis. Postmodernist feminist theory has been properly critical of feminist theory that has attempted to transpose what has been the experience of some women into generalized formulations legitimized with reference to the authenticity of women’s expe- rience. At the same time s/he may refashion it to meet what s/he is trying to get said and this refashioning is projected into the future. (Smith 1987: 160–161) They are the expert practitioners of their own lives and the ethnographer’s work learns from them and goes beyond what they know only in finding how they are connected beyond what is visible to them. ignores that extraordinary moment earlier in the women’s movement when we (women) were discovering how to think ourselves apart from masculinized forms of thought. particular place. we turn to how Mikhail Bakhtin conceives of speech genre. Any par- ticular utterance (particular speaker. It is essentially spoken or written and hence is necessarily and always mediated by the structuring of the discourse in which it is uttered. the perspectives and views [for those of the people she talks to or observes]. positioned. but as a dialogue. Each next speaker or writer is in dialogue with discourse. particular time. Experience The concept as it is used here does not refer to an authentic individual act of consciousness that gives access to a world directly known. . The critique asserts that there is no experience that is not already dis- cursively structured. Her/his utterance is determined by what is already established. Bakhtin’s theory is dialogic. but not subjects. We had no language for speaking our condition. Not as an already given authentic knowledge of what is for the individual consciousness. a con- cept closely parallel to that of discourse. however. particu- lar others) draws on what has been laid down in the past to get something said and at the same time launches the genre into the future with whatever modi- fications have been introduced to realize the speaker’s intentions. using its terminology and syntactic strategies s/he reproduces the genre. Speech genres are laid down as people speak and write with others in the settings of their activities. This is how I conceive of experience. This critique. The latter constituted us as other.

INSTITUTIONAL ETHNOGRAPHY 43 now an actual conversation. We might imagine institutions as nodes or knots in the [ruling] relations . people are active in producing the generalized out of the particular. it is dialogically engaged with what the speaker or writer is trying to find a way of telling. titles. We speak from the known-in-practice ongoing concerting of actual activities. And are general- ized. Institution As in institutional ethnography: the best I can do here is to quote from an earlier work in which I first wrote about institutional ethnography: I am using the terms ‘institutional’ and ‘institution’ to identify a complex of rela- tions forming part of the ruling [relations]. Discourse does not determine. We speak methodi- cally. … the social organization of our daily practices governs our choice of syntactic forms and terms when we speak them. It is like this: while we may be sure that we under- stand everything that is said. organized around a distinctive func- tion – education. who may question. we typically find that there is a problem of speech. but rather the intersection and coordination of more than one relational mode of [ruling]. and so on. it has definite conventions. ‘institution’ does not identify a determinate form of social organiza- tion. speak of that social organization. names. We speak knowing how rather than knowing that. When we first encounter a new social organizational setting. laws. we are not at all sure that we will be able to speak cor- rectly. Sometimes there is a struggle with the discourse within which the dialogue is engaged to get it to speak what is coming to be experience-as-uttered. coordinating multiple strands of action into a functional complex. Hence in institutional settings. and both are interpenetrated by relations of discourse of more than one order. One speaks of her/his experience and the other. Integral to the coordinating process are ideologies systematically developed to provide categories and concepts expressing the relation[ship] of local courses of action to the institutional function. we are not sure that the appropriate terms.. assigns the privilege of narration to the person whose experience is uttered. We do not. health care. … It would be hard for someone to speak unmethodically in referencing social organization of which she is a competent practitioner. In contrast to such concepts as bureaucracy. whether in writing to an abstract reader or spoken directly to another. As a social form. except in rare instances. Language and social organization Again.. (Smith 1987: 160) It is a specific capacity of institutions that they generalize. and so forth will fall . of course. I quote: As interviewers we persuade people to talk about the everyday worlds in which they are active. Characteristically state agencies [and laws] are tied in with professional forms of organization.

but registers as an underlying determinant of how the informant talks of the setting because it is the only way in which it makes sense to talk. or documents related to a case. it is the case file. Some of them are critical to women’s safety. regulations. Interchanges are connected through routing instructions and procedures. Almost all interchanges are structured by the required use of forms. There are a few unfortunate golf courses that bar the foot traveller and in one or two spots. But what are these appropriate forms? They are seldom as distinct as matters of protocol. In the inter- view situation. They are commonly the difference between how those speak who are ongoing practitioners of a world and who know how to use its language in situ as part of its ongoing concerting and how those speak who are as yet feeling their way into the properties of its everyday practices. the original setting is not operative. the clerk. or laws which screen. prioritize. Toronto is riddled with ravines. (Pence 1996: 55) Ravines in Toronto A surprising city. reports. No single person hand-carries it from one processing point to another. and filter the information the worker uses to produce accounts. administrative procedures.44 PUTTING THE PRACTICE INTO THEORY into place or – what is perhaps even more important – that we will know how to assemble these in the appropriate ways. is the first person in a long chain of responders to a domestic assault case. original emphasis) Processing interchanges No one oversees a case from its inception to its final resolution. others do not. This file stands in for the woman who was assaulted. for her assailant. (Pence 1996: 53) Processing interchanges are designed to organize the information received by intervening practitioners and to institutionally structure the kind of information that is produced at each interchange. the only way through is the railroad line and its insecurities. The case is routed. It is neither the worker nor the woman who was beaten who moves from one point to the next in the stages of case processing. Toronto’s ravines play a part in Margaret Atwood’s The Edible Woman. and for those who act to inter- vene. The language of the setting observes the rela- tions of its social organization. Its proper uses indeed preserve them. but generally it is possible to travel substantial dis- tances with only occasional returns to the surface to find a connecting link between one part of a ravine and another where the path has been closed off by narrowing sides. many of which have not been developed. the ravines are modestly but not too enthusiastically groomed. (Smith 1987: 188–189. in some parts they remain quite wild and even occasionally difficult. The dispatcher. . shape. of different official language uses. or in civil court. The station of each respon- der has built into its information-collecting and information-producing functions mechanisms that link the information into an overall case construction. Some of these connections operate quite smoothly. Trails have been created.

Institutions are specific functional foci within the complex of ruling relations. drawn or otherwise reproduced. parent–child. and the like are rethought as relations among people that rely on the materiality of the text and its increasingly complex technologies. discourse (in the Foucauldian sense). Such concepts as information. We know them variously as bureaucracy. In contrast to those forms of coordinating people’s activities that are direct (moving furniture. ‘culture’. Smith 1990b: 93–96. the stability and replicability of organi- zation or institution.E. elec- tronic and so on) of what is written. computer. Social relations ‘Social relations’ does not refer to relationships such as instructor–student. science. . rather. institutions.or trans-local and based in texts of various kinds (print. film. or playing soccer). A social ‘relation is not a thing to be looked for in carrying out research. mass media. Texts Printed or electronic or otherwise replicable texts have the extraordinary capacity of double presence: they are read or produced in the actual local settings of people’s work or other activities and at the same time their repli- cation in multiple different settings (and at different times) enters a stan- dardized component into every settings in which the same text is read/viewed. and so on). INSTITUTIONAL ETHNOGRAPHY 45 Ruling relations The ruling relations become visible when we start enquiry in the everyday worlds of people’s experience. television.W. Smith 1995: 24). for example. It is a practical realization of the commitment to the discursive problematic of the everyday/everynight world (D. it is what is used to do the looking’ (G. film. man- agement. Rather it directs attention to and takes up analytically how what people are doing and experiencing in a given local site is hooked into sequences of action implicating and coordinating multiple local sites where others are active (D. they are extra. and so on and so on. Unlike some theorizing of ‘text’. They are relations that coordinate people’s activities across and beyond local sites of everyday experience. and so on. knowledge. Materiality is emphasized because we can then see how they are used to cre- ate a crucial join between the everyday actualities of people’s activities and the social relations they coordinate. between lovers. in and out of the ephemer- ality of people’s everyday activities. the term is used here strictly to identify texts as material in a form that enables replication (paper/print. G.W. It is. Smith 1987). Smith 1995). I suggest.E. texts that produce.

For wages-for-housework theorists. between work and leisure. that they mean to do. Competence shifts the ground away from the concerting of people’s activities and towards an individuating of the social. Feeding the family.46 PUTTING THE PRACTICE INTO THEORY Work Dialogues with experience can participate in many discourses. being described as consumption or not at all. The merit of this concept when inquirers enter into dialogue with their respondents is that it bridges both subjective and objective. let alone as housework. doesn’t just mean preparing the food and putting it on the table. As George Herbert Mead (1947) said in recommending his brand of behaviourism over that of Watson. the personal. the act begins inside the individual and proceeds outside and inside and outside cannot be separated. which is based upon work as paid employment and does not apply to housework. and so forth. Here is how I conceptualized this in my original formulation of institutional ethnography. (Smith 1987: 165) The concept of work as used in institutional ethnography identifies what people do that takes some effort and time. it also . We are familiar with the way in which the concept of work had not been extended in the past to women’s work in the home as housewife. eating lunch in the cafeteria or making and eating sandwiches. This generous conception of [work] includes not only domestic labor proper but such activities as driving to one’s place of employment. pur- chasing and maintaining clothes. though the institutional ethnographer relies on the knowhow of her or his respondents. as Marjorie DeVault (1991) has described it. in relation to whom. They have used it to include all the work done by women [as well as by men] to sustain and service their … functioning in the wage relation and hence indirectly to sustain and service the enterprises employing their labor. that relies on definite resources. Some wages-for-housework theorists have developed an expanded concept of housework. under what conditions. The dis- course of institutional ethnography selects ‘work’ as a metaphor that focuses the examination of experience on what people do. which I shall use as a model. It is both about what people do and it is also about the consciousness that necessarily goes along with doing. Our notion of work had to be expanded to include housework. and in doing so we discovered some of its pre- suppositions – the implicit contrast. and is organized to coordinate in some way with the work of others similarly defined. It installs people’s doing in a dis- position and formulated thus the social never actually happens. Often it’s thought that experience must refer to the subjective. and with what resources. All these aspects of everyday life are essential to the economy though they would not ordinarily be described as work. Expanding the concept of work for our purposes requires its remaking in more ample and generous forms. The concept of work as it is used here to focus the attention of both parties to the dialogue on what is done and being done. housework becomes an economic category identifying those work processes that are in fact part of the economy but are not represented as work. for example. I do not want this notion to be shifted towards the notion of competence.

in M.W. thinking through what might already be in store somewhere. M. because our time was now more restricted.W. Austin. Speech Genres and Other Late Essays. McGee. Austin. This inside/outside disclosure anchors the ethnography in an everyday/everynight world while avoiding the divorce of subjective from objective that characterizes research oriented to ‘meaning’. (1995) ‘Beyond the ruling category to what actually happens: notes on James Mill’s historiography in “The history of British India”’. M. Bakhtin. and so on. London: Virago. TX: University of Texas Press. NOTE 1 We started interviewing in one community doing interviews that sometimes lasted a couple of hours and met two or three times. INSTITUTIONAL ETHNOGRAPHY 47 means planning the meal. (1986a) Speech Genres and Other Late Essays.M. Arnup. in M. translated by V. . Princeton. All research has an economy that shapes it beyond methodology and theory. K. (1981) The Dialogic Imagination: Four Essays.M. McGee. (1986b) ‘The problem of speech genres’. (1998) Remaking Women: Feminism and Modernity in the Middle East. We had to move to another community and there. REFERENCES AND BIBLIOGRAPHY OF INSTITUTIONAL ETHNOGRAPHIC STUDIES Abu-Lughod. translated by V. (1994) Education for Motherhood: Advice for Mothers in Twentieth-Century Canada. is to be able to extend this concept beyond what people might ordinarily think of as ‘work’ and beyond how we might ordinarily think about it. coordinating the cooking and serving of a meal with the diverse schedules of different family members. Bannerji. Atwood. It also anchors the terminology in the discourse of the respondent’s setting which is shared by others in the same setting and bears the social organization of their coordinated work. we worked with a more structured interview procedure. edited by M. Campbell and A. deciding what has to be shopped for and how to fit the shopping into the daily routine.M. On the other hand. H. At some point it may hamper rather than help and then sociologists need to remember that they are focused on people’s doing in the most general sense and need to find some other concept. Bakhtin.M. M. (1997) The Edible Woman. Bakhtin. L. NJ: Princeton University Press. The trick. Toronto: University of Toronto Press. Experience. Toronto: University of Toronto Press. M. Manicom (eds) Knowledge. however. we were rudely ejected by the superintendent of schools. Holquist. Austin. Bakhtin. TX: University of Texas Press. TX: University of Texas Press. We had made our contacts through the schools and after only two or three interviews and one meeting with a principal. and Ruling Relations: Studies in the Social Organization of Knowledge. it’s important too to keep in mind that ‘work’ here is a metaphor. to see thinking or reading as work or the production of objectivity in science as people’s con- certed work.

Thousand Oaks. Berkeley. DeVault. and Gereffi. Holstein (eds) Handbook of Interviewing. power’. M. Philadelphia. Toronto: University of Toronto Press. and Ruling Relations: Studies in the Social Organization of Knowledge. (1994) ‘U. Dickinson (eds) Health Illness and Health Care in Canada. Campbell. Campbell. Gurstein (eds) Learning to Write: Women’s Studies in Development. T. Chinchilla.J. (1998a) ‘Institutional ethnography and experience as data’. Campbell and A.J. Experience. (1990) The Logic of Practice. Walker and J. R. Studies in Political Economy 27: 29–51. October. Campbell. M. (eds) (1986) Writing Culture: The Poetics and Politics of Ethnography. A. University of Toronto. York University. and Ruling Relations: Studies in the Social Organization of Knowledge. (1986) ‘Social policy and everyday life in nursing homes: a critical ethnography’. D. in M. in E. Clifford. A.L. CA: Sage.A.295. (1998b) ‘Research on health care experiences of people with disabili- ties: exploring the everyday problematic of service delivery’. N. Campbell and A. DeVault. M. L. experience. Campbell./S. (1998) ‘Engendering development theory from the standpoint of women. (1992) Making Gray Gold: Narratives of Nursing Home Care. Diamond. New York: Harcourt Brace. in D. (1984) ‘Information systems and management of hospital nursing: a study in the social organization of knowledge’. Toronto: University of Toronto Press. (1988) ‘Management as ruling: a class phenomenon in nursing’. Qualitative Sociology 21(1): 55–73. (1995) ‘Teaching accountability: what counts as nursing education?’ in M. and Ruling Relations: Studies in the Social Organization of Knowledge. 2nd edn. Gubrium and J.A. M. Experience.M. Toronto. M. De Montigny. paper presented at ‘Exploring the Restructuring and Transformation of Institutional Processes: Applications of Institutional Ethnography’. N. L. Manicom. M. Bullock. Currie. CA: Stanford University Press. and Marcus. Toronto: Garamond Press. PA: Temple University Press. Diamond. (eds) (1995) Knowledge. (1990) ‘Community care: ideology and experience’. P. in B. L. Ng. Bonacich. Stanford. Singh Bolaria and H.287–1.F.L. .L. G. Experience.L. A. (1995b) ‘The power of being professional’. Campbell. G. M. Currie. Hamilton and P. Campbell. and Wickramasinghe. G. M. in J. PhD thesis. Ong (eds) Global Production: The Apparel Industry in the Pacific Rim. Chicago: University of Chicago Press. and McCoy. Knowledge. Cheng. J.A. CA: University of California Press. Manicom (eds) Knowledge. Campbell. (1994) ‘The structure of stress in nurses’ work’. (1995) ‘Literacy.E.48 PUTTING THE PRACTICE INTO THEORY Bourdieu.D. G. G. retailers and Asian garment production’. (1995a) Social Working: An Ethnography of Front-line Practice. Toronto: University of Toronto Press. Manicom (eds) Knowledge. (1991) Feeding the Family. (2000) ‘Institutional ethnography: using interviews to investigate ruling relations’. De Montigny. Jacob (eds) Community Organization and the Canadian State. and Manicom. M. and Ruling Relations: Studies in the Social Organization of Knowledge. Vancouver: Collective Press. in M. in R. Chicago: University of Chicago Press. Cheng. Toronto: University of Toronto Press. Darville. G. Experience. T. Noga and P. Campbell and A. Toronto: University of Toronto Press. Social Science and Medicine 23(12): 1.

(1995) ‘Mothering. (1995) ‘Compulsory heterosexuality: schools and lesbian students’. Garfinkel. Foucault. Experience. Knowledge. schooling and the family wage’. Foucault. Boulder. education. A. 1972–1977. Kinsman. Foucault. schooling. P. and curriculum reform’. Griffith. Campbell and A. Hull. vol. K. Griffith. text. M. institutions. and children’s development’. in J. (1995) ‘“These things just happen”: talk. A.I. G. D. D. Knowledge. Experience. M. London: Tavistock. Manicom. in M. G. A. L. CT: JAI Press. Gadamer. New York: Continuum. Griffith. Foucault.E.I. PhD dissertation. in G. Greenwich. Manicom (eds) Knowledge. schooling and social class’.R. . Jackson. Chicago: University of Chicago Press. and Ruling Relations: Explorations in the Social Organization of Knowledge. in M. Campbell and A. (1984) ‘Ideology. York University. Miller and J. (1979) Discipline and Punish: The Birth of the Prison. M.. University of Toronto.L. London: Routledge. Griffith. and Smith.) Untying the Text: A Poststructuralist Reader. Toronto: University of Toronto Press. R. (1995) ‘Foreword’. Toronto: University of Toronto Press. Manicom (eds) Knowledge. and Ruling Relations: Studies in the Social Organization of Knowledge. 2.M. Human Studies 21(4): 347–360. A. M. CO: Westview. (1970) The Order of Things: An Archaeology of the Human Sciences. Campbell and A. R. Campbell and A. and Smith. Griffith. H-G. A. NJ: Prentice-Hall. community and Asian immigrant women’s work: a study in labor market organization’. and the social organization of knowledge’. (1972) The Archaeology of Knowledge and the Discourse on Language. Holstein (eds) Perspectives on Social Problems.. (1987) ‘Constructing cultural knowledge: mothering as discourse’. paper presented at ‘Exploring the Restructuring and Transformation of Institutional Processes: Applications of Institutional Ethnography’. N. D.I. (1967) Studies in Ethnomethodology.I. C. in M. and single parent families: the normative ordering of families through schooling’. (1996) The New Work Order: Behind the Language of the New Capitalism. and Ruling Relations: Studies in the Social Organization of Knowledge. Gaskell and A. Foucault. (1995) ‘The textual practices of sexual rule: sexual policing of gay men’. H. (1995) Writing Ethnographic Fieldnotes. (1994) Truth and Method. Gee. Griffith. in M. (1999) ‘State. Toronto: University of Toronto Press. Khayatt.E.E. Toronto. McLaren (eds) Women and Education: A Canadian Perspective. 2. PhD dissertation. and Ruling Relations: Studies in the Social Organization of Knowledge. A. Manicom (eds) Knowledge. Holstein (eds) Perspectives on Social Problems. Manicom (eds) Experience. New York: Pantheon. and Shaw. Toronto: University of Toronto Press. (1998) ‘Educational restructuring in Ontario’. Experience.L. INSTITUTIONAL ETHNOGRAPHY 49 Emerson. October. University of Toronto. New York: Pantheon. Campbell and A. vol. (1998) ‘Ethnography. in G.I. D. Heap. (1980) Power/Knowledge: Selected Interviews and Other Writings.I. Grahame. Greenwich. in M. (1981) ‘The order of discourse’. Miller and J. Fretz. J. in R. Englewood Cliffs. Experience. and Lankshear.I. CT: JAI Press. M. Grahame. New York: Vintage. Toronto: University of Toronto Press. (1990a) ‘Coordinating the uncoordinated: mothering. and Ruling Relations: Studies in the Social Organization of Knowledge. and Smith. Calgary: Detselig. Young (ed. (1990b) ‘What did you do in school today? Mothering. J.

C. (1988) ‘Constituting class relations: the social organization of teachers’ work’. York University. Experience. McCoy. (1995) ‘Activating the photographic text’. (1998) ‘We do meaning: health services research. Experience. and Ruling Relations: Studies in the Social Organization of Knowledge. L. Toronto: University of Toronto Press. Campbell and A. Chicago: University of Chicago Press. in M. Campbell and A.W. G. Toronto: University of Toronto Press. (1999) ‘Accounting discourse and textual practices of ruling: a study of institutional transformation and restructuring in higher education’. Mykhaloviskiy. and Ruling Relations: Studies in the Social Organization of Knowledge. Experience. Manicom (eds) Knowledge. in M. PhD thesis. Manicom. E. in M. University of Toronto. A. and Sachs. Mueller.50 PUTTING THE PRACTICE INTO THEORY McCoy. R. A. N. Ng. October. Self and Society: From the Perspective of a Social Behaviorist. A. Oxford: Basil Blackwell. Cultures and Organizations. Pence. unpublished doctoral dissertation. paper presented at ‘Exploring the Restructuring and Transformation of Institutional Processes: Applications of Institutional Ethnography’. (1995) ‘What’s health got to do with it? Class. Studies in Societies. Pence. F. (1986) Politics of Community Services: Immigrant Women. edited by C. Toronto: Garamond Press. unpub- lished doctoral dissertation. Toronto. Toronto. A. Ng. (1976) The German Ideology. Manicom (eds) Knowledge. (1990) ‘Immigrant women: the construction of a labour market category’. Manicom (eds) Knowledge. paper pre- sented at ‘Exploring the Restructuring and Transformation of Institutional Processes: Applications of Institutional Ethnography’. A. October. research transfer and the social organization of interpretive practices’. D. Manicom (eds) Knowledge.H. University of Toronto. Ng. R. Mead. Ng. R. . Macdonnell. (2001) ‘Safety for battered women in a textually-mediate legal system’. Morris. Mueller. (1986) Theories of Discourse: An Introduction. and Engels. (1991) In and Against Development: Feminists Confront Development on its Own Ground. (1987) ‘Peasants and professionals: the social organization of women in development knowledge’. Mueller. (1996) ‘Safety for battered women in a textually-mediated legal system’. University of Toronto. East Lansing. (1947) Mind. MI: Michigan State University. Manicom. Toronto: University of Toronto Press. labour adjustment and the transformation of garment workers’ lives’. R. Naples. Experience. University of Toronto. K. Marx. and Ruling Relations: Studies in the Social Organization of Knowledge. (1995) ‘Multiculturalism as ideology: a textual analysis’. in M. Campbell and A. Canadian Journal of Women and the Law 4: 96–112. York University. Campbell and A. Moscow: Progress. Toronto: University of Toronto Press. Class and State. gender and teacher’s work’. L. (1995) ‘Beginning in the standpoint of women: an investigation of the gap between “Cholas” and “Women of Peru”’. Rural Sociology 65(2): 194–214. (1998) ‘Immigration policy. and Ruling Relations: Studies in the Social Organization of Knowledge. E. E. unpublished doctoral dissertation. (2000) ‘Standpoint epistemologies and the uses of self- reflection in feminist ethnography: lessons for rural sociology’. special issue on ‘Institutional Ethnography’.


Rankin, J.M. (1998) ‘Health care reform and restructuring of nursing in British
Columbia’, paper presented at ‘Exploring the Restructuring and Transfor-
mation of Institutional Processes: Applications of Institutional Ethnography’,
York University, Toronto.
Rankin, J. (2001) ‘Texts in action: how nurses are doing the fiscal work of health
care reform’, Studies in Societies, Cultures and Organizations, special issue on
‘Institutional Ethnography’.
Reimer, M. (1988) ‘The social organization of the labour process: a case study of
the documentary management of clerical labour in the public service’, PhD
dissertation, University of Toronto.
Reimer, M. (1995) ‘Downgrading clerical work in a textually mediated labour
process’, in M. Campbell and A. Manicom (eds) Knowledge, Experience, and
Ruling Relations: Studies in the Social Organization of Knowledge. Toronto:
University of Toronto Press.
Schutz, A. (1962) ‘On multiple realities’, in Collected Papers, vol. I. The Hague:
Martinus Nijhoff.
Sèchehaye, M. and Renée, P. (1969) ‘Autobiography of a Schizophrenic Girl, with Analytic
Interpretation, translated by G. Rubin-Rabson. New York: Grune and Stratton.
Sharma, N. (1998) ‘Capitalist restructuring and the social organization of
“non-immigrants” in Canada: the reformalization of non-membership’, paper
presented at ‘Exploring the Restructuring and Transformation of Institutional
Processes: Applications of Institutional Ethnography’, York University, Toronto,
Simpson, P.E. (1998) ‘An inquiry into the Canadian development worker role as
a site for the reorganization of Canada’s international relations’, paper
presented at ‘Exploring the Restructuring and Transformation of Institutional
Processes: Applications of Institutional Ethnography’, York University,
Toronto, October.
Smith, D.E. (1974) ‘Women’s perspective as a radical critique of sociology’,
Sociological Inquiry 4(1): 1–13 (a revised version was published in D.E. Smith
Smith, D.E. (1977) ‘Some implications of a sociology for women’, in N. Glazer
and H. Waehrer (eds) Women in a Man-Made World, revised edn. Chicago:
Smith, D.E. (1979) ‘A sociology for women’, in J. Sherman and F. Beck (eds) The
Prism of Sex: Essays in the Sociology of Knowledge. Madison, WI: University of
Wisconsin Press.
Smith, D.E. (1987) The Everyday World as Problematic: A Feminist Sociology. Boston,
MA: Northeastern University Press.
Smith, D.E. (1990a) The Conceptual Practices of Power: A Feminist Sociology of
Knowledge. Toronto: University of Toronto Press.
Smith, D.E. (1990b) Texts, Facts, and Femininity: Exploring the Relations of Ruling.
New York: Routledge.
Smith, D.E. (1997) ‘The underside of schooling: restructuring, privatization, and
women’s unpaid work’, Journal for a Just and Caring Education Fall: 14–29.
Smith, D.E. (1999) Writing the Social: Critique, Theory and Investigations. Toronto:
University of Toronto Press.
Smith, G.W. (1988) ‘Occupation and skill: government discourse as problematic’,
occasional paper no. 2, The Nexus Project: Studies in the Job Education Nexus,


Smith, G.W. (1990a) ‘Policing the gay community’, in R. Ng, G. Walker and
J. Muller (eds) Community Organization and the Canadian State. Toronto:
Garamond Press.
Smith, G.W. (1990b) ‘Political activist as ethnographer’, Social Problems 37:
Smith, G.W. (1995) ‘Accessing treatments: managing the AIDS epidemic in
Ontario’, in M. Campbell and A. Manicom (eds) Knowledge, Experience, and
Ruling Relations: Studies in the Social Organization of Knowledge. Toronto:
University of Toronto Press.
Smith, G.W. (1998) ‘The ideology of “fag”: barriers to education for gay students’,
Sociological Quarterly 39(2): 309–335.
Stock, A. (2000) ‘An ethnography of assessment in elementary schools’, doctoral
dissertation, University of Toronto.
Turner, S.M. (1995) ‘Rendering the site developable: textual organization in the
planning process’, in M. Campbell and A. Manicom (eds) Experience, Knowledge,
and Ruling Relations: Explorations in the Social Organization of Knowledge. Toronto:
University of Toronto Press.
Ueda, Y. (1995) ‘Corporate wives: gendered education of their children’, in
M. Campbell and A. Manicom (eds) Knowledge, Experience, and Ruling Relations:
Studies in the Social Organization of Knowledge. Toronto: University of Toronto
Walker, G. (1990a) Family Violence and the Women’s Movement: The Conceptual
Politics of Struggle. Toronto: University of Toronto Press.
Walker, G. (1990b) ‘The conceptual politics of struggle: wife battering, the
women’s movement and the state’, Studies in Political Economy 33: 63–90.
Walker, G. (1995) ‘Violence and the relations of ruling: lessons from the battered
women’s movement’, in M. Campbell and A. Manicom (eds) Knowledge,
Experience, and Ruling Relations: Studies in the Social Organization of Knowledge.
Toronto: University of Toronto Press.
Walker, G. (1998) ‘Internal pressures, external forces: the ongoing struggle of the
battered women’s movement and the state’, paper presented at ‘Exploring the
Restructuring and Transformation of Institutional Processes: Applications of
Institutional Ethnography’, York University, Toronto, October.


Sam Porter

This chapter is about how ethnographers might deal with the issue of social
structures. Two main questions are addressed. First, are there such things as
social structures? Second (predicated upon a positive answer to the first
question), is it possible to use ethnographic methods to examine those struc-
tures and their relationship with social action? The suggested answer to this
is that the philosophy of science known as critical realism provides a sound
basis for moving ethnography beyond the examination of specific social
instances, in order to examine the general structural context of those
The chapter has three main sections. The first section entails an historical
review of the way (primarily sociological) ethnographers have dealt with
the issue of structure. It starts by looking at the foundational work of
Durkheim and Weber, and their attempts to incorporate the structuring of
human action into their theories of research. Following from this baseline,
there is an examination of the phenomenological and postmodernist turns
in ethnographic theorizing, which concentrates on how the influence of
these two movements led to the erasure of structure from the ethnographic
imagination. The section concludes with a brief review of what Brewer
(2000) terms post-postmodernist approaches to ethnography, which is seeking
tentatively to pull back from the extremes of methodological individualism
and scepticism.
The second section introduces critical realism and attempts to argue how
the use of this philosophical position as a grounding for ethnography is
capable of solving many of the problems raised by phenomenology and
postmodernism in relation to the place of structure in our understanding of
the social world.
The final section works through a practical example of the use of critical
realist ethnography. This is an ethnographic study of power relations
between nurses and doctors in an intensive care unit which attempts to gen-
eralize beyond the interactions observed and comments of individuals
heard, in order to show how those relations are socially structured.



The elementary forms

Probably the most seminal attempt to understand the nature of structures
using ethnographic information was that of Durkheim’s Elementary Forms of
Religious Life (1995 [1912]), which relied for empirical data upon ethno-
graphic studies of the religious practices of Australian aboriginals. However,
the point of Durkheim’s use of ethnography was not simply to describe the
ideographic experiences of the individuals observed. His aim was far bolder:

I will make every effort to describe the organization of this system with all the care
and precision that an ethnographer or a historian would bring to the task. But my
task will not stop at description. Sociology sets itself different problems from those
of history or ethnography. It does not seek to become acquainted with bygone
forms of civilization for the sole purpose of being acquainted with and recon-
structing them. Instead, like any positive science, its purpose above all is to explain
a present reality that is near to us and thus capable of affecting our ideas and
actions. That reality is man. . . . I have made a very archaic religion the subject of
my research because it seems better suited than any other to help us comprehend
the religious nature of man, that is, to reveal a fundamental and permanent aspect
of humanity. (Durkheim 1995: 1)

The latter sentence does not entail a theological statement, rather a funda-
mentally sociological one. As Durkheim stated elsewhere:

Religion contains in itself from the very beginning, even in an indistinct state, all
the elements which, in dissociating themselves from it, articulating themselves,
and combining with one another in a thousand ways, have given rise to the various
manifestations of collective life. (Durkheim 1897, cited by LaCapra 1985: 247)

Thus, Durkheim’s project was to use ethnographic material on the religious
life of an ‘organic’ society (cf. Durkheim (1984 [1893]), where, ‘[a]t the same
time all is uniform, all is simple’ (Durkheim 1995: 5), in order to uncover the
elementary forms of religion which constituted one of the most fundamen-
tal structures underpinning collective life. By examining the undifferentiated
components, as displayed under conditions of organic solidarity, Durkheim
then argued that the significance of this ‘fundamental and permanent aspect
of humanity’ could be traced through the development of society, allowing
us to understand better how religion, or at least the transformed remnants of
religion, continued to have a profound influence upon the collective life of
modern society.
The purpose of this brief synopsis is not to defend Durkheim’s thesis on
the social structural position of religion, still less to defend his belief in the
veracity of positive science. Rather, it is to note that from the era of its for-
mulation as a distinct academic discipline, sociology approached ethno-
graphy from a particular standpoint. Namely, it aimed to use ethnographic


material to tell us something wider about social life than the particular
experiences of those who were the subject of ethnographic studies.

Weber’s Verstehende

It might be argued that contemporary ethnography has an altogether differ-
ent sociological parentage, namely the Verstehende sociology of Max Weber,
which was based on a distinctly nominalist position, and thus would seem
to place the focus of research firmly on individuals rather than structures:

When reference is made in a sociological context to a state, a nation, a corporation,
a family, or an army corps, or to similar collectivities, what is meant is ... only a
certain kind of development of actual or possible social actions of individual persons.
(Weber 1978 [1956]: 14, original emphasis)

The problem with such a position is that, whether or not it is correct, people
in general tend not to adopt it, and instead regard collectivities as having a
reality beyond that of individuals, and act accordingly. This is a point that
Weber concedes, but to which he responds with an idealist interpretation:

These concepts of collective entities … have a meaning in the minds of individual
persons, partly as of something actually existing, partly as something with norma-
tive authority. .. . Actors thus in part orient their action to them, and in this role
such ideas have a powerful, often decisive, causal influence on the course of action
of real individuals. (Weber 1978 [1956]: 14, added emphasis)

Thus, Weber is brought to a position whereby, while denying the material
reality of structures, he accepts that they have real causal effects. In doing so,
he comes very close to Durkheim’s position, who, while arguing for the coer-
cive power of social facts, is perfectly prepared to accept that they ‘have no
existence save in and through individual consciousness’ (Durkheim 1982
[1895]: 52). The significant issue here is that Weber’s acceptance of the causal
efficacy of social collectivities leads him further to accept that research focus-
ing on the meanings of individuals is not sufficient on its own to gain full
sociological understanding.

A correct causal interpretation of typical action means that the process which is
claimed to be typical is shown to be both adequately grasped at the level of mean-
ing and at the same time the interpretation is to some degree causally adequate. ...
even the most perfect adequacy on the level of meaning has causal significance
from a sociological point of view only insofar as there is some kind of proof for the
existence of a probability that action in fact normally takes the course which has
been held to be meaningful. (Weber 1978 [1956]: 12)

There are ambiguities here in relation to the distinction between cause and
correlation, ambiguities which were played out in Weber’s own empirical


analysis of the relationship between religion and socio-economic organization
(see Parkin 1982). However, the point is that, like Durkheim, Weber had
aspirations for sociological knowledge that went beyond the ideographic
uncovering of understandings of particular individuals.

Phenomenological ethnography

While Durkheim and Weber made significant statements about the use of
ethnography as a method of sociological exploration, they were less than
comprehensive on ontological, epistemological and methodological details
of their approaches. Despite the considerable length of Elementary Forms, it
is almost entirely taken up with empirical information and theoretical
extrapolations based on that information. Durkheim pays little heed to the
methodological quandaries relating to ethnography that contemporary
ethnographers would take as standard; questions such as the status of
accounts of western ethnographers examining non-western social practices,
or of the secondary use of ethnographies. While Weber’s development of the
Verstehende method was more comprehensive, he remained vague about key
conceptual issues. This vagueness was seized upon by Alfred Schutz in his
Phenomenology of the Social World (1972):

The present study is based on an intensive concern of many years’ duration with
the theoretical writings of Max Weber. During this time I became convinced that
while Weber’s approach was correct and that he had determined conclusively the
proper starting point of the philosophy of the social sciences, nevertheless his
analyses did not go deeply enough to lay the foundations on which alone many
important problems of the human sciences could be solved. Above all, Weber’s
central concept of subjective meaning calls for a thoroughgoing analysis. As Weber
left this concept, it was little more than a heading for a number of important pro-
blems which he did not examine in detail. (Schutz 1972: xxvii)

Schutz’s argument that Weber failed to state clearly the essential attributes
of Verstehen, subjective meaning or social action allowed him the opportu-
nity to develop Weber’s methodology in his own direction. Specifically, it
allowed Schutz to graft on his own interpretation of Husserlian phenome-
nology. The key notion that he adopted from Husserl was that what appears
to be the natural ordering of the world is in fact the result of conceptual
judgements of the mind.
This move meant that Weber’s rather ambiguous, but essentially two-way
conception of the causal flow between structures and actions was now
firmly grounded in a uni-directional causal flow, with structures being
reduced to the status of epiphenomena of subjectivities. For Schutz, the
social world was ‘essentially only something dependent upon and still
within the operating intentionality of ego-consciousness’ (1967: 44). The
problems of this emphasis on the subjective at the cost of recognition of the
causal effects of the wider social world upon the subjectivities of individuals


have been identified by Giddens (1976), who has criticized Schutz’s position
on the grounds that it can take account neither of the unacknowledged
effects of actions, nor of determining conditions that are not mediated by the
consciousness of actors.
This discussion of Schutz has been confined to the level of theory. We need
to ask what are the consequences of such a position being adopted by the
empirical ethnographer? I wish to argue that the significance of Schutz’s
ideas for ethnographic sociological research lay in the tendency for many
ethnographers to rely exclusively on uncovering, in an unproblematic fashion,
the subjective interpretations of individuals, at the cost of examining how
social structures and processes influenced those interpretations.
Let me take one example, that of Hockey’s (1986) ethnographic study of
British soldiers. Hockey clearly situates his study within the subjectivist tradi-
tion, noting that his research involved stressing ‘the importance of interpret-
ing the behaviour of people in terms of their subjectively intended meanings’
(1986: 10). Part of the study involves an account of soldiers on combat duty in
rural Ireland. In good ethnographic tradition, the soldiers’ everyday lives are
richly described. Unfortunately, description is as far as it goes. The simple but
crucial question of why the soldiers were acting in the way that they were in
the location in which they found themselves is beyond the remit of this sub-
jectivist inquiry. For example, we are given a description of highly armed
troops, who are psychologically on a combat footing in a foreign country, the
native inhabitants of which they treat with, at best, suspicion. This information
is of limited utility unless we ask why this is so, and in order to do so, we need
to understand how the political, social and economic relations between
Ireland and Britain have been historically structured.
The restriction of the interpretation of behaviour to the subjectively
intended meanings that immediately generated it obviates the possibility of
deeper analysis of the social situation encountered by the ethnographer. In
short, the social phenomenological assumption that individual interactions
and interpretations are all there are leads to analytic superficiality. While
understanding the interpretations of the social actors is a necessary condition
for sociological knowledge, it is not a sufficient one (Porter 1993).
The import of this critique of phenomenological ethnography is that if
ethnography is to be an effective method of social research, it needs to be
grounded in an ontological, epistemological and methodological model that
can provide a deeper understanding than subjectivism is capable of; one
which is able to link the subjective understandings of individuals with the
structural positions within which those individuals are located.

Postmodernist ethnography

As we have seen, Durkheim, in his study of elementary forms of religious
life, felt few qualms about using ethnographic data about the beliefs and
practices of people in a culture radically different to his and using them to


make sweeping generalizations about the nature of human society in toto.
Thus, the intricacies and complexities of the lives of these people were sub-
sumed into the framework of his functionalist ‘meta-narrative’, to use
Lyotard’s (1984) term.
Here we can see the ethnographer as a figure of authority, claiming the
right to explain people’s lives from his or her singular point of view. This
issue of explanatory presumptuousness was especially acute in the disci-
pline of anthropology, where one found western ethnographers presuming
to explain various non-western cultures according to their own western
lights. This academic imperialism involved an uncomfortably close affinity
with the military, economic and political imperialism of western capital
and nations.
The construction of the non-western ‘other’ within the self-serving rubric
of the western paradigm has come in for severe criticism (see, for example,
Said 1978). One of the results of this sort of critique has been a crisis in con-
fidence on the part of western ethnographers. One of the responses to this
crisis has been to radically undermine the authority of the ethnographic
author. This involves moving from a critique of the particular meta-narratives
that ethnographers adopted to frame their explanations to a generalized
attack on the use of any sort of meta-narrative. Thus, for example, Crapanzano
warns that the events being described by the ethnographer will be ‘sub-
verted by the transcending stories in which they are cast’ (1986: 76). Here we
can see a full swing across the spectrum of epistemological confidence –
from the point where ethnographers assume unproblematically the validity
of their authorial position, to the point where ethnographies are seen as
nothing more than the inventions of their authors (see Clifford 1986).
A concomitant rejection of patriarchal meta-narratives led many feminist
ethnographers to adopt a similar critique. Thus, for example, Stanley (1987)
developed her radically perspectivist ‘reverse archaeology’. Reverse archaeo-
logy goes beyond the traditional perspectivist tenet that different observers
will have different perspectives on the same phenomenon, to asserting that
the same observer will have constantly different perspectives on the social
patterns that they observe. Once again, we have moved from the unreflexive
certainties of patriarchal thought to the acceptance of the indeterminacy of
From particular politico-epistemological problematics, postmodernism
rapidly moved to occupy a central space in thinking about qualitative
research (see, for example, Scheurich 1997; Cheek 2000). In short, pluralism,
perspectivism and scepticism became the order of the day.
The rejection of authorial meta-narrational certainties and their replacement
by a robust epistemological scepticism has even more profound conse-
quences for our knowledge of the social world than that of the phenomeno-
logical turn in qualitative research. Phenomenology still held on to the anchor
of subjective experience, and the possibility that the perspective of the subject
was amenable to interpretation at the point of encounter between its horizon
and that of the researcher’s (see Gadamer 1975). With postmodernism, even

Similarly. (1998) state that ‘Our research suggests the persistence of signifi- cant structural limitations in the pathways to adulthood. for what they say can claim no superiority in terms of adequacy over that which anyone else says. that which looks at the health behaviours of young people. The difficulty with such a position is that. this does not obviate the need for some sort of generalization. To take one substantive area of research. we can glean the importance of understanding subjective meanings as the basis of social action. Thus. whereby it is claimed that aspects of a particular situation can be seen to be instances of a broader recognizable set of features. it is increasingly being recognized that we cannot understand why young people behave as they do in relation to the consumption of such things as tobacco and alcohol unless we include social structure in our explanation. it has to posit some form of generalizable truth claims. it is done so at the cost of abandoning the very raison d’être of ethnographic research. in that if ethnography is to be of any utility for our understanding of society. is the reductio ad absurdum of the postmodernist position. for example. there are indications that post- modernism has passed its high-water mark. I believe. There is now a growing accep- tance that an adequate ethnographic model needs to incorporate these insights. Thus. As a consequence. This. if ethnographies are simply authorial inventions. There are good practical reasons for this. in that it is they who are required to demonstrate the broad pertinence of their work.412). If absolute uncertainty and rela- tivism are accepted. we are made aware of the dangers of making absolute claims about those understandings. Post-postmodernist ethnography While it is as yet too early to tell definitively. Malcolm Williams argues that while strong claims about the capacity to generalize from ethnographic data cannot be sustained. while at the same time going beyond them. This is not to say that the issues highlighted by both phenomenology and postmod- ernism can be rejected out of hand. then what is the point of ethnography? While the problem of intellectual arrogance is solved. rather than reflections. These moves beyond postmodernism are not confined to the realms of sociological theorizing. of greater or lesser accuracy. From phenomenology. of social reality. there is little else for ethnographers to say about the social world. Denscombe (2001) puts forward the thesis that one of the primary . From postmodernism. CRITICAL REALIST ETHNOGRAPHY 59 this is abandoned to the kaleidoscope of changing patterns and perspectives that allow us little or no confidence to assume that one interpretation of the social world can claim epistemological superiority over any other. he advocates ‘moderatum generalizations’. indeed they are far more crucial to the endeavours of practising qualitative researchers. although these may be different to those experienced by previous generations’ (1998: 1. Pavis et al. in order to take into account the patterning of social behaviour. in Chapter 5 in this book.

Already. it would be a model which accepts that there is a reality beyond individuals. while at the same time avoiding the errors in previous models identified by phe- nomenology and postmodernism. As such. in order for events to occur as they do. it might be argued that this transcendental conjuring up of structures is just another example of the arrogance of modernist meta-narration. Such a model would be required to allow for the possibility of generalization. From a postmodernist point of view. Denscombe wishes to uncover from whence such uncertainty comes. because actions are patterned. its claims are contentious. In other words. Thus. Examples from other substantive areas of qualitative research could be used just as well to demonstrate that while researchers remain sensitized to postmodernism’s attack on the arrogance of sureness. it asks what factors lead to the patterning of human understand- ings and actions. rather than random. The assertion that structures exist is as far as critical realist a priori reasoning goes – identification of . A number of attempts have been made to construct just such a realist basis for ethnography. Critical realism differs from both of these models in that it gives less away to the phenomenological and postmodernist critiques. they must be structured in some way. but which does not over-extend its claims about how much we can know about that reality (in response to postmodernism) or about the degree to which external reality controls the decisions of individuals (in response to phenomenology). The first stage of this argument is to outline what the philo- sophical position of critical realism entails and how it relates to ethnography. a priori. In short. This leads to the question of what sort of philosophy of science can be used to ground this post-postmodernist (Brewer 2000) position. In response it is important to underline the limits of this transcendental move. in relation to society. there is a drawing back from its extreme position. The a priori answer to the question is that. However. While the notion of uncertain identity may seem in itself to be a very postmodern one. He concludes that ‘the evidence from this research would still tend to support the contention … that the social/historical context of late modernity heightens the significance of uncertain identities in the lives of young people’ (2001: 175). what is needed is a realism that is not naive. and an acceptance that it is both permissible and desirable to make knowledge claims which extend beyond the individual.60 PUTTING THE PRACTICE INTO THEORY factors in explaining adolescent smoking are the ‘uncertain identities’ of young people. which asks what must be the case. we have got to the stage where a number of possible objections need to be answered. I hope to argue that the use of critical realism as a basis for ethnography is both viable and useful. most notably Hammersley’s (1992) ‘subtle realism’ and Altheide and Johnson’s (1998) ‘analytical realism’. THE TENETS OF CRITICAL REALISM At the basis of critical realism is an acceptance of the crucial importance of what Kant (1896) termed the transcendental question.

First. for the most part unconsciously repro- duce (or occasionally. critical realism argues that the perceptual criterion of reality is not the only criterion. except in the minds of individuals. in their conscious human activity. The argument here is that if structures do not exist as per- ceivable things. Thus. critical realism argues that this does not warrant the further phe- nomenological assumption that activity can be solely explained in terms of individual consciousness. for example. To treat human beings as objects whose actions are determined by extraneous factors is to misinterpret the nature of humanity. but discovering whether or not this is the case is an empirical issue. There are two related reasons for the rejection of individualist reductionism. But it is nevertheless the unintended consequence (and inexorable result) of. and thus consists of fallible knowledge claims. in certain cases. and it does so because it is grounded in the meanings of individual actors. belief in the reality of social structures is justified on the grounds of the perceivable effects they have upon patterns of human behaviour. critical realism might be accused to the error of reification. (Bhaskar 1989b: 80) Social phenomenology’s insistence that the ethnographer should concen- trate exclusively on understanding the subjective meanings and actions of . then they do not exist. In response. From a Weberian point of view. Similarly. which turns on the capacity of an entity to bring about changes in material things. the wider social effects of actions may not be those consciously intended by the actor. This is the reason why positivist social scientific attempts to identify universal empirical regularities have come to naught. Thus. On this crite- rion. their activity. Neither of these forces can be directly perceived. or work to reproduce the capitalist economy. but their effects upon material things can be. This is not to say that. CRITICAL REALIST ETHNOGRAPHY 61 the nature of those structures is a matter of empirical endeavour. the social context within which an indivi- dual lives provides the conditions for her consciousness. Second. economic factors may indeed be the primary structural influence upon the patterning of human relations. as it is also the necessary condition for. Entities in this category include gravity and magnetism. Thus people do not marry to reproduce the nuclear family. and adds to it the causal criterion. to be is not to be perceived. transform) the structures that govern their substantive activ- ities of production. for example. arguing that human behaviour displays con- siderable variation. but to be able to do (Bhaskar 1989a). because we can see its effects upon the patterning of iron filings. transcendental real- ism does not warrant the sort of vulgar Marxist argument concerning the deterministic role of economic relations. While accepting the phenomenological point that human activity is con- scious. Thus Bhaskar (1989b) characterizes the relationship between society and people thus: people. The phenomenologist in turn might point to the apparent structural determinism of such a position. we know that the structural mechanism of magnetism exists.

Thus. Critical realism seeks to overcome this dichotomy through its adoption of a modified naturalism. so the tendencies they ground may not be invariant over space and time (Bhaskar 1989a). they are only relatively enduring. It is noted that within the pos- itivist paradigm. this does not involve the adoption of the traditional struc- tural functionalist line. But of what do the structures which generate tendencies consist? The criti- cal realist answer to this question is ‘structures of relations’. Thus. On the one hand. they do not exist independently of social actors’ conceptions of what they are doing as they act. However. in order to exclude confounding variables. social structures do not exist independently of the activities they govern. which asserts that the laws operat- ing in the social world are essentially of the same quality as those operating in the natural world and can therefore be studied in the same manner (see Comte 1974). the search for invariant cause and effect relations is. On the other hand. unlike natural structures. thus requiring radically different methods of study (see Winch 1959). The initial purpose of social science thus becomes one of uncovering how these relationships are structured. positivism adheres to naturalism. conducted within the closed system of experimental design. Third. which in both cases involves the move from observed pheno- mena to generative structure (Bhaskar 1989a). while both the natural and social worlds are to be understood through elucidation of the relationship between structure and phenomenon. and it is the specific combination of structural influ- ences that will determine the nature of events. critical realists have abandoned constant conjunction as the bedrock for the identification of causal structures. Instead. the approach taken mirrors closely that of Marx: ‘Society does not consist of individuals. the relations within which these individuals stand’ (Marx 1973 [1858]: 265). The commonality between the natural and social sciences lies in the connection between empirical investigation and theory construction. along with Marx. critical realism’s relational model of society argues that social relations are emergent from and irreducible to indi- viduals. the nature of that relation- ship is crucially different. critical realism seeks to go further . wher- ever possible. I describe the uncovering of structural relations as the initial purpose of social science because. They do not pertain because events are rarely subject to a single generative structure. However. In contrast to methodological individualist models. This is not to say that critical realism conceives of natural structures as determining constant conjunctions of events. or a worker because of one’s relationship with an employer. the hermeneutic tradition sees the social world and the natural world as radically different arenas. First. This ignores the fact that both the natural and social worlds are open systems. but expresses the sum of interrelations. which sees groups as having a fundamental reality that can provide the bedrock of social explanation (Collier 1994).62 PUTTING THE PRACTICE INTO THEORY individuals is at least partially grounded in the traditional dichotomy in social science between the positivist and hermeneutic traditions. replacing it with the notion of tendencies. Second. where constant conjunctions do not pertain. one is only a wife by dint of one’s relationship with a husband.

In fact. structure and action are seen as distinct but interdependent. while ‘I got married because I love him’ or ‘I took a job with Siemens because computers are the big thing in the twenty-first century’ could be seen as examples of lack of consciousness. media. the role of social science is not only to uncover the structuring of relations. However. often reproduce oppressive structures. (Bhaskar 1989b: 3–4) As is indicated in the above quotation. for their survival. it examines the structuring of human relations using the criterion of whether they promote or constrain the human freedom and dig- nity of those involved in them. thus society does not exist independently of human agency – the error of reification. original emphasis). oppressive structures depend upon actions based upon false consciousness. because struc- tures are dependent for their existence upon actions. Such a model contains within it the currently rather unfashionable notion of false consciousness. A similar analysis could be made of the effects of capitalist relationships on workers’ lives. We do not create society – the error of voluntarism. People. rules and resources for everything we do. such as getting married or getting a job. Instead. through their actions. It is the unmotivated condition for all our motivated productions. Here we can see a dialectical relationship between structure and action – oppressive structures promote the adoption of false consciousness by individuals. Society provides the means. But these structures which pre-exist us are only reproduced or transformed in our everyday activities. while the actions predicated upon that false consciousness are functional to the continued sur- vival of those structures. to use the examples cited above. Thus. The social world is repro- duced and transformed in daily life. .. it does not base that rejec- tion upon the collectivist reification of social entities adopted in positivist and functionalist interpretations of society: the existence of social structure is a necessary condition for any human activity. critical realism rejects the fact–value distinction. Thus. in various ways. CRITICAL REALIST ETHNOGRAPHY 63 than this: ‘The philosophers have only interpreted the world.. the point is to change it’ (Marx and Engels 1970 [1847]: 123. In a similar fashion to critical theory. the structured relations involved in the patriarchal nuclear family can be sub- jected to critical analysis in order to ascertain the degree to which they con- strain the human freedom and dignity of female partners within those relationships. ‘A woman’s place is in the home’ or ‘wages are fair exchange for labour’ might be seen as exam- ples of false consciousness. but also to use that information as the basis for informed action to remove the sources of structural oppression. Their distinctness is grounded in temporal . The relationship between structure and action It can be seen from our examination of critical realist philosophy that it adopts a position that rejects the individualist voluntarism at the core of both phenomenology and postmodernism. or at least upon lack of consciousness of their oppressive effects.

For meanings cannot be measured.64 PUTTING THE PRACTICE INTO THEORY differences – the social structures in which individuals are located pre-exist those individuals. it is now necessary to examine how the ethnographic method might fit into such a model. rather than being one. not everyone in society acts in the same way and there will often be groups in society who benefit from the status quo. but are rather the result of earlier social actions. namely to overcome the dichotomy between interpretivism and structuralism. However. only . from a critical realist perspective. As a result. As Bhaskar puts it: the conceptual aspect of the subject-matter of the social sciences circumscribes the possibility of meaningful measurement. While critical realism accepts Giddens’s point that structures cannot exist independently of social actions. . structure and action exist only in the same instance. it asks the ques- tion. The place of ethnography in critical realist social science Having outlined the general tenets of critical realism’s views of social science. in that it distinguishes critical realism from another influential conception of the relationship between structure and action – Giddens’s (1984) theory of structuration. the crucial difference lies in his conception of the relationship between structure and action. Structures enjoy only a virtual reality until they are ‘instantiated’ through social actions.. the structural conditions pertaining will often place limitations on the pace of change. This factor is significant. it is argued that structuration theory ‘cannot recognize that structure and agency work on different time intervals’ (Archer 1995: 89). However. In other words. It will be remembered that one of the limitations of naturalism accepted by critical realism centred on the observation that social structures do not exist independently of social actors’ conceptions of what they are doing when they act. ‘whose actions?’ and answers ‘the actions of the past’. Archer (1995) uses the example of the literacy campaign in revolutionary Cuba. Thus. and who will therefore have the motivation to resist change. Second. structure and action are simply two sides of the one coin (Craib 1992). The social struc- tures that affect our lives have not been conjured up by our own instantiation of them. transformation of structures through action is rarely an instantaneous process. In that case.. What contemporary actions do is either maintain or transform already developed structures. There are two main reasons for this. in that it points to the limits of the use of quantification in the social sciences. First. Acceptance of this facet of the relationship between social action and social structure has considerable epistemological consequences. Giddens’s project is similar to that of critical realism. For Giddens. the structural conditions of the educational system inherited from the Batista regime meant that there simply were not enough people equipped to teach literacy to make alterations to that structure anywhere near instantaneous. structure and action have distinct realities of their own which operate on different timescales.

The advantages of such an approach is that it enables research to bring social science back into the ethnographic equation. Hypotheses about them must be expressed in language. racist beliefs and acts cannot be explained solely through elucidation of the attitudes of the individuals involved. and to point to the sort of actions required to make them less oppressive. And precision of meaning now assumes the place of accuracy in measurement as the a posteriori arbiter of theory. to show how these relations may be oppressive. AN EXAMPLE OF CRITICAL REALIST ETHNOGRAPHY Thus far. can be categorized as a social structure. the discussion has been somewhat esoteric. Thus. I wish to provide an example to show how critical realist ethnography might work in practice. it is used as a method to uncover the manifest interactions of the social world. original emphasis) Because the subject matter of social science is conceptual. In this section. racist- engendered positions predate actors currently situated within them. Moreover. but to use examination of those understandings and actions as part of the process of uncovering the relationship between agency and structure. The most significant issue to note was my acceptance that racism. Second. it abandons many of the methodological assumptions normally associated with ethnography (cf. Because such an . and confirmed in dialogue. and beyond that. the purpose of ethnographic investigation is not to ideographically illumi- nate the understandings and actions of individuals. First. because it involves enduring relations between actors in different social positions. Looked at another way. Rather than confining its focus to individual experience. or rejecting the notion of knowledge altogether. which are then subjected to the transcendental process of theory generation to infer the structural conditioning of those interactions. However. I should elucidate the assumptions upon which I was working. It is for this reason that ethno- graphic methods have a role to play in critical realist social science. Being Black or Asian in a White racist society often entails involvement in enduring disadvantageous relations. Brewer 1991). there is a need for the qualitative testing of theories about the nature and effects of social struc- tures upon social actions and vice versa. Language here stands to the conceptual aspect of social science as geometry stands to physics. while critical realism continues to use ethnographic tech- niques of data collection. Before recounting the ethnographic data gleaned in this research. (Bhaskar 1989a: 46. In other words. The example I wish to use involves my examination of the rela- tionships pertaining between Black and Asian doctors and White nurses in the intensive care unit of an Irish hospital (Porter 1993). the role of ethnography is twofold. it is used to subsequently test the veracity of theories concerning the nature and effects of the structures pertaining. CRITICAL REALIST ETHNOGRAPHY 65 understood. critical realism is able to use ethnographic data to illuminate structured relations.

Nonetheless. rules and resources available to enable or coerce action. by providing the means. reprimanded them for their behav- iour. Thus individuals will enjoy more or less powerful positions in a racist relationship depending on how they are categorized in racist terms. there will be pressure to do so. In order to avoid such a static conception of racism. those structures. Individuals’ actions are not determined by the structural relations within which they live. This is not to say that the existence of structural racism means that all interactions between White and Black or Asian individuals will necessarily be characterized by racism. that social structures are only relatively enduring and dependent upon the conceptions of actors. and a fortiori in relation to human actions. and openly criticized their professional competence. media. It will be remembered that this notion of con- stant conjunction is rejected by critical realism both on the grounds that the relationship between generative structure and event takes place in an open system. This was not the interpretation that Hughes. directed doctors to perform tasks. will engender ten- dencies towards certain courses of actions. These reasons were synthesized in the statement that ‘differential competence in utilizing relevant bodies of social knowledge is perhaps the most salient interactional manifestation of “status” characteristics’ (Hughes 1988: 18). The interaction between White nurses and Black or Asian doctors showed little alteration in the balance of power as compared to relationships between White nurses . Second. acts and structured rela- tions needs to be considered. the articulation between racist attitudes. that the doctors studied were unfamiliar with British culture and thus dependent upon nurses for cultural translation. in that racist attitudes can provide an ideological ratio- nalization for the structural inequalities from which a White person may benefit. favoured. He posited two reasons for nurses’ lack of deference. that relations between young inexperienced Asian doctors and mature White nurses led to dilemmas in status. utiliz- ing an interactionist model. it leads to their reification (Rex 1970). Using the transcendental move. The ethnographic observations from my study elicited considerably different characteristics from those found by Hughes (1988). In relation to racism in the specific context of hospitals. First. White nurses frequently abandoned outward shows of defer- ence. I had the benefit of previous ethnographic work in a British casualty unit which had discov- ered that a significant variable affecting relationships between nurses and doctors was geographical origin (Hughes 1988). allowing the nurses to become far more actively involved in therapy than they would otherwise have been. Hughes discovered that. Rather. in their interaction with doctors who had recently immigrated from the Asian subcontinent. Being White in a racist society does not nec- essarily mean that one will adhere to racist ideology. I concluded that the patterns of behaviour and attitude displayed by White nurses in Hughes’s (1988) ethnography took the configurations that they did because they were influenced by the gener- ative structure of racism.66 PUTTING THE PRACTICE INTO THEORY individualist approach is unable to explain the origins of individual attitudes.

[The staff nurse reddens but does not reply. Indeed. He wryly concluded: ‘but I think people soon get the message’. this doctor was used by nurses as a font of knowledge. This dressing should have been changed hours ago. When ‘backstage’ (Goffman 1969). a complaint about the authoritarianism of another doctor was couched in the following terms: . rather. the consultant comes across a heavily stained dressing. or. However. There’s a lot of exudate and we’re having to change the padding almost continually. [One con he identified was] that every time he got a new post. Is he pyrexial? This sort of authoritarianism was rarely openly challenged by nurses.] Sister: The wound’s been giving us a lot of trouble. some nurse–nurse inter- actions in my study revealed a different story. out of earshot of doctors and patients. the doctor most respected for his clinical knowledge was African. However. A subsequent informal interview with the doctor confirmed that this flaunting of knowledge was part of a deliberate strategy on his part: I asked him what it was like coming to work in a place like this. For example. Consultant: Hmm. Consultant: This is a disgrace. an Asian consultant had a reputation for distant superiority. though their relationship with this doctor was one of starched compliance rather than cooperation. Every time. Another strategy used by Black and Asian doctors was authoritarianism. On occasion these complaints were observed to be expressed in racist terms. he had to go through the same old routine before they accepted that he was good at his job. observations of his interactions with nurses indi- cated that he made conspicuous efforts to ensure that nurses fully appreciated just how much he knew. which involved minimal communication with nurses beyond the issuing of instruc- tions or castigations: In the process of a ‘ward round’. With a background in medical research. Thus. He went over some of the pros and cons of his move. CRITICAL REALIST ETHNOGRAPHY 67 and doctors. in the unit observed. Thus. nurses often used the opportu- nity to complain to each other about the behaviour of their medical col- leagues. he had to start all over again. encounters between doctors and nurses observed in my ethnographic study displayed considerably more deference than those encountered in Hughes (1988). Consultant: Who is looking after this patient? Staff nurse: I am. just how little the nurses knew in compar- ison to him. People automatically assumed that because he ‘wasn’t from here’ he wouldn’t know anything.

while the agents involved thought and acted within the matrix of a racist generative structure. interactions observed by Hughes (1988) were characterized by significant alterations in power relations. but where there was little interactional evi- dence of a significant alteration in power relations between Black and Asian doctors and White nurses. For example. Yet such an explanation. Moreover. as the experiences of second and third generation Black British people demonstrates (Brown 1984). It is my contention that critical realist ethnography can provide a fruitful alter- native to these dichotomous approaches. there are dangers in making too close a linkage between status and cultural competence. it might seem that Hughes’s interactionist interpretation is justified on the grounds of its parsimoniousness. for the most part. Someone should tell her we’re not Indian peasants. lived in Ireland for a considerable time and were familiar with its culture. My data could be accom- modated in this explanation by noting that status differentials remained latent precisely because there was no differential competence in the utilization of social knowledge. where racist attitudes were occasionally privately expressed. and were therefore dependent on White nurses for cultural translation. the African clinical expert was described thus: Staff nurse: He’s the smartest Black person I’ve ever met. status differentials were manifested in differential competence in the utilization of social knowledge. For Hughes. Nor was racism simply a pathological reaction to heavy-handed authoritar- ianism. From comparison of the data. but was on occasion adopted in relation to those doctors who were not regarded by nurses as autocratic. in that it can explain the data without recourse to the transcendental reference to social structures. by concentrating on manifest inter- actions. To summarize and compare the ethnographic data recounted thus far. We are thus in a situation where an interactionist approach is unable to explain racism as an enduring set of social relations. but where the variable manifestations of overt racism observed belie crude structuralist explanations. because that structure operated within an open system along with other structures. It is certainly not the case that racism is suffered exclusively by those unattuned to British or Irish culture. does not address the issue of why ‘dilemmas in status’ between Black and Asian and White members of society exist in the first place. obser- vations in my study uncovered a latent racism. Another difference between the two studies was that while Hughes reported that the Asian doctors observed tended to be unfamiliar with the cultural cues of British culture. its . In contrast.68 PUTTING THE PRACTICE INTO THEORY Staff nurse: I think she thinks that that dot on her head makes her a maharaja or something. Critical realism allows us to accept that. the doctors in my study had.

it wouldn’t be proper. the professional power entailed in being a member of the medical profession tends to enhance it. Func- tional specificity leads doctors to be judged solely according to the level of their medical skills. Acceptance of universalistic- achievement values means doctors are seen as having attained their occupa- tional status solely through their ability. this does not mean that racism ceases to exist. It would be a bit unprofessional … Staff nurse 1: It’s not exactly professional to get down on your knees in the middle of the clinical room … Staff nurse 2: Yea. CRITICAL REALIST ETHNOGRAPHY 69 effects were seen as tendencies. We have already seen how nurses’ acceptance of professional ideology . From such a perspective. but you couldn’t say ‘Listen you. To be acceptable. Criticisms of cultural incompetence can be dressed up as universalistic. naked racism. Given the acceptance of these values as criteria for judging professional actions. in his list of pattern variables. affective and diffuse. being particularistic. Professional power depends upon the acceptance by others of the ideology of professionalism. achievement- oriented. Yet. The professional pattern variables put forward by Parsons are largely accepted within the culture of nursing. Because nurses hanker after a professional status that they have not fully attained and which they see so suc- cessfully adopted by medicine. you’re not in Arab land now’. the racism is still there. In cases where a doctor’s cultural familiarity is such that this vehicle cannot be used. ascriptive. Affective neutrality assumes that judgements should be made solely on the basis of scientific rationality. Ethnographic evidence that this tension between racist and professional ideologies was experienced by some nurses comes from the following discussion between two nurses concerning a Muslim doctor carrying out his religious observances in the unit: Staff nurse 1: Why didn’t you say something to him about it? Staff nurse 2: Well. Another structure which was influencing the nature of agents’ thoughts and actions was that of occupational relations animated by professional power. is not a justifiable form of social interaction between nurses and doctors. criticisms of cultural com- petence can be seen as vehicles through which racism is expressed. which has probably been most clearly (if naively) identified by Parsons (1951). I suppose I could have told him that clinical areas have to be open to staff at all times. It is not just that social structures operate in an open system. They are also dependent for their maintenance or transformation on the actions of indivi- duals. they take professional ideology very seriously indeed. affectively neutral and functionally specific to the skills of medicine. rather than constant conjunctions. racism needs to take on a ‘rational’ form that does not appear to flout the tenets of professional ideology. While being Black or Asian in a racist society tends to reduce one’s social status. but that an avenue for its expres- sion in the public arena is closed off. This is what is happen- ing when differential cultural competence is being married to ‘race’.

it is incumbent upon ethnographers to take into account the fact that the social world is a complex and often contradic- tory place. Elucidation of such a multilayered social situation requires at least two things from the researcher. to use Bourdieu’s (1990) term – which both promote and are promoted by per- taining structures. Furthermore. . volitional agents. An example here would be the motivations underpinning the actions adopted by doctors to reinforce their professional status and thereby limit the possibility of racist interactions. but non-determining social relations. A more subtle and successful strategy was adopted by the African doctor who portrayed himself as the font of medical knowledge. it was accepted with bad grace. we need to recognize that gaining knowledge about that world is a far from unpro- blematic process. This in turn requires acceptance of the existence of structured. it requires theoretical work to explain why individuals’ interactions take the patterns that are observed using those techniques. SUMMARY From the social context examined. Such a pic- ture can emerge only through the utilization of close observational techniques. Because this strategy was more closely attuned to the assumptions of professional ideology. Second. First of all. While this authority was rarely openly challenged by nurses. We have noted two strategies. Moreover. While structures provide the context for actions. we can see the complex relationship that exists between social actions and structures. it requires a clear picture of the inter- actions of individuals. One was the use of authoritarianism which relied rather crudely on the occupational status of the doctor.70 PUTTING THE PRACTICE INTO THEORY has the effect of maintaining the professional power of doctors. both at the level of action and of motivation. his status was both formally and conatively accepted by nurses. This acceptance was deliberately fostered by the actions of a number of Black and Asian doctors. it has no option but to claim the capacity to provide generalizable knowledge. that does not mean that we need to abandon the notion of the social either to the methodological individualism of phenome- nology or to the anti-realist scepticism of postmodernism. If ethnography has any pretensions to providing information that will be of practical use. that context is itself determined by the particular combi- nation of structures pertaining and the degree to which actions reinforce or undermine the relative strength of those structures. actions may be the result of taken-for-granted assumptions – or doxic experience. CONCLUSIONS Following the debunking of traditional social scientific verities by pheno- menology and postmodernism. Or actions may be the result of agents’ reflexively considered strategies. Here I am thinking of nurses’ adherence to professional ideology. inhabited by thinking. However.

in J. in J. edited by A. in turn. M. A. J. E. M. Crapanzano. Lincoln (eds) Collecting and Interpreting Qualitative Materials. A. (1989b) Reclaiming Reality: A Critical Introduction to Contemporary Philosophy. Craib. London: Sage. J. Cambridge: Polity. J. J. Bourdieu. There would therefore seem to be a significant dissonance between theory and practice. Hence my contention that the marriage of critical realist theory with ethnographic methods holds out such fruitful possibilities. structured. London: Routledge. At the same time. P. Clifford. M. Bhaskar. I. London: Verso. Marcus (eds) Writing Culture: The Poetics and Politics of Ethnography. (1994) Critical Realism: An Introduction to Roy Bhaskar’s Philosophy. (1984) Black and White in Britain. London: Croom Helm. Brewer. CA: University of California Press. while it is rarely acknowledged. (2001) ‘Uncertain identities and health-risking behaviour: the case of young people and smoking in late modernity’. in N.E. the presentation of generalized claims is a common fea- ture of contemporary ethnographies. at least to some degree.E. Cambridge: Cambridge University Press. Collier. individualism and scepticism remain powerful forces within qualitative methodological theoriz- ing. Unless ethnographers are pre- pared to settle for claims that their work entails an activity akin to fiction writing. R. (1989a) The Possibility of Naturalism: A Philosophical Critique of the Contemporary Human Sciences. Archer. Berkeley. Buckingham: Open University Press. (1992) Anthony Giddens. London: Sage. and Johnson. (2000) Postmodern and Poststructural Approaches to Nursing Research. . Andreski. then there is a need to return to the epistemology of generalization and the ontology of structure. at best. (1990) The Logic of Practice. CRITICAL REALIST ETHNOGRAPHY 71 That claim. Denscombe. as Malcolm Williams notes (Chapter 5 in this volume). Clifford and G. London: Verso. (1991) Inside the RUC: Routine Policing in a Divided Society. Bhaskar. (2000) Ethnography. (1966 [1897]) Suicide. Oxford: Clarendon. (1974) The Essential Comte. V. Brewer. Comte. Berkeley. Durkheim. Marcus (eds) Writing Culture: The Poetics and Politics of Ethnography. D. It has been the contention of the chapter that critical realism can be used to facilitate just such a return. Clifford and G. Brown. 2nd edn. New York: Free Press. R. or. Cheek. (1986) ‘Hermes’ dilemma: the masking of subversion in ethno- graphic description’. (1986) ‘Introduction: partial truths’. one-off portraiture. C. rests upon the assumption that human activities and relations are. (1998) ‘Criteria for assessing interpretive validity in qualitative research’. Indeed. CA: University of California Press. London: Harvester Wheatsheaf. London: Heinemann. This is not a sustainable position. REFERENCES Altheide. (1995) Realist Social Theory: The Morphogenetic Approach. Denzin and Y. British Journal of Sociology 52(1): 157–177.

translated by W. Kolegar. IL: Free Press. Müller. J.E. E. D. Women’s Studies International Forum 10(1): 19–31. (1970 [1847]) The German Ideology. E. Gadamer. Cunningham-Burley. (1959) The Idea of Social Science. (1998) ‘Health related behavioural change in context: young people in transition’. I. (1896) Critique of Pure Reason. (1951) The Social System. (1988) ‘When nurse knows best: some aspects of nurse/doctor inter- action in a casualty department’. Harmondsworth: Penguin. Shils and C. translated by F. Halls. Harmondsworth: Penguin. A. edited by G. London: Macmillan. (1993) ‘Critical realist ethnography: the case of racism and professional- ism in a medical setting’. M. Lyotard.418. (1984) The Postmodern Condition: A Report on Knowledge. Hammersley. Sociology 27(4): 591–609. Durkheim. (1986) Squaddies: Portrait of a Subculture. Roth and C. Kant. New York: Seabury. Rex. H-G. T. LaCapra. Pavis. (1969) The Presentation of Self in Everyday Life. (1995 [1912]) The Elementary Forms of Religious Life. M. Winch. London: Weidenfeld and Nicolson. Rheinstein. Cambridge: Polity. (1982) Max Weber. London: Routledge and Kegan Paul.72 PUTTING THE PRACTICE INTO THEORY Durkheim. London: Heinemann. (1973 [1858]) Grundrisse. and Amos. A. Weber. (1985) Emile Durkheim: Sociologist and Philosopher. (1975 [1960]) Truth and Method. New York: Pantheon. (1985 [1893]) The Division of Labor in Society. Porter. (1997) Research Method in the Postmodern. Social Science and Medicine 47(10): 1. Glencoe. A. H. L. (1972) The Phenomenology of the Social World. J. F. Wittich.407–1. London: Hutchinson. (1978 [1956]) Economy and Society. (1984) The Constitution of Society. Glyn-Doepel. Volume 1. Marx. K. G. Gerth. Goffman. (1978) Orientalism. E. London: Falmer. Sociology of Health and Illness 10(1): 1–22. (1976) New Rules of Sociological Method. Durkheim. Schutz. Scheurich. New York: Free Press. Stanley. Giddens. E. Said. F. Henderson. London: Macmillan. S. (1992) What’s Wrong with Ethnography? London: Routledge. Schutz. London: Ellis Horwood/Tavistock. K. Exeter: University of Exeter Press. and Engels. Marx. translated by W. translated by W.D. Wittich. A. C. E. Hughes. E. Halls. P. The Hague: Nijhoff. S. Fischoff. T. (1970) Race Relations and Sociological Theory. F. A. Manchester: University of Manchester. Parkin. S. Chicago: University of Chicago Press. . Berkeley. Fields.D. D. (1990) Reflections on the Making of an Ethnographic Text. New York: Free Press. Giddens. (1987) ‘Biography as microscope or kaleidoscope? The case of “power” in Hannah Culwick’s relationship with Arthur Munby’. M. CA: University of California Press.. Manchester: Manchester University Press. Williams. (1967) Collected Papers. (1982 [1895]) The Rules of Sociological Method. Parsons. A. translated by E. J-F. Hockey. translated by K. Wright Mills. J. A. London: Lawrence and Wishart. Roth. Parsons.

Recent theorizing glossed as postmodernism (Poster 1990. a governing mode of irrationality. into a ground. These turns on rationality leave little room for focused and context-based fieldwork that begins with a context within which deciding takes place. Rickert. while fashioning versions of Durkheim. I present an alternative focus using framing. Rationality. working in the idealist tradition. THE RATIONAL Parsons’s magisterial work. presumes an analytic metalanguage that recasts the concrete. This chapter traces the most important consequence of such movements of thought in the emergent role of rationality in qualitative sociology. that surface belies depth. They have built on this synthesis. from which occasioned and occasional rationality emerges. This turn. Weber). the grasped. which serves those in power. Marshall and Pareto. drawing upon German idealism (Kant. Dilthey. rioters.3 FRAMING THE RATIONAL IN FIELDWORK Peter K. It is argued that the situational rationality of detective work is ill suited to easy rationalization. Parsons. The synthesis of rationality performed by Weber and recast by Parsons remains the sociological touchstone. and murderers. showing the workings of such mannered rationality. This writing transforms rationality from a constant ‘figure’ of modernity. emphasized the value-bases for the three types of social action identified by . The Structure of Social Action (1936). Lemert 1995) reverses the enlightenment tradition reflected in sociology. established a working agenda for an entire generation of sociologists. In many ways. the field and the surround. Following an overview of some of the ways in which modern social science has grappled with rationality. and characterize homicide detective work as situ- ated and situational rationality.1 The implications of this formulation for qualitative work are examined in a final section. is in fact patterned rationality. technology and modern civic administration. science. as something quite different. are mis- leading and false or at best factious. A number of interpretive approaches consistent with qualitative work are discussed below. understood best against irrationality as a ‘ground’ in which the occasional appearance of rational action is notable. what passes for the rational is merely the conventional wisdom. Manning Sociological analysis rests fundamentally on irony: demonstrating that what appears is not what is. that the apparent irra- tionality of criminals.

and their tensions underlie change. Weber delineated individual irrationality.g. Espeland 1998. meant many things. ‘Rational’ in his vocabu- lary pointed to diverse processes. as in organizing a gallery exhibi- tion. and are in tension with forces of institutionalized rationality. Concrete forms of rationality. with diverse his- tories exist. sentiments and impulses of people. doctrine. ‘objective’ rationality associated with institutionalized procedures. The rational in the Weberian scheme did not stand alone. transcendental values and patterns of choice. The underlying issue was comparative analytic sociology – how can generalizations about social order be produced given the tensions between forms of rationality. Merelman 1998). Similarly. Many kinds of rationality. Weber knew of Nietzsche’s profound distrust of the modern rational state and its suppression of passion. antinomial movements and nihilism. between situated deciding and the long- term legitimized rationality. This formulation not only means that competition emerges between forms of rationality within organizations. culture and personal autonomy. norms and meaning was always embedded in broader questions . feelings. The interplay of values. guidelines. courts. saw that given a commit- ment to. The broader context of rationality grounds the specific instances and concrete examples of it. rational and charismatic were bases of legitimization for subjective meanings attached to rules and commands. religions. bureau and other forms of public discourse and administration. and the forces of irrationality that constantly arose from cults. the offence. individual subjectivities and collective rationality? The diverse meanings of rationality were for Weber displayed in the ten- sions between culture. and these procedural devices were replicated in the courts. can decide cases using concrete. precedents. both substantive and formal rationality. for example. substantive rationality based on the offender. from the cumulative. As Antonio (1995) elegantly demonstrates. it was based on the notion that means and ends were connected within sets of elaborate rules and offices that constitute modern bureaux. Traditional.74 PUTTING THE PRACTICE INTO THEORY Weber. one could act to sustain such values rationally. the passions. but also suggests tensions between objec- tive and substantive rationality. while adhering formally to stipulated procedures. the victim or its symbolic importance (e. seen as a process of associating ends and means. science. furthermore. and resistance to them shape modern social life. this instrumental rationality was invisibly grounded in other assumptions about social life such as Protestant aestheticism. This Parsonian synthesis obscured Weber’s very complex and subtle posi- tion (see Sica 1988. Weber. It was set always in the context of deciding. rules and evidence. and signified complex historical developments and substantive content. The idea of rationality stood on the methodical linking of efficient means to well-accepted ends. enlightenment values and a rational politics (Merelman 1998: 351). politics and the state administrative structure. This ideal-typical construction set the stage for modern studies of rationality. an aesthetic cluster of values such as modern art. a ‘hate crime’) rather than on formal rationality based on legal procedures. In many respects.

rules some matters. requires gift exchanges that function to integrate and bind parties. or formal rationality. Furthermore. are ‘surface’ features. They have elevated substantive rationality to demonstrate the perverse rationality of even the most exotic. As a result. treatment of infectious and many chronic diseases (Fabrega 1998). renews belief in norms of religious piety. forgiveness and obligation. This system remains a cultural gloss because it explores the powerful and consistent consequences of a unified culture rather than of individual deciding. Symbolic anthropology Social scientists influenced by Clifford Geertz (1973) ground their arguments in detailed explications of context such that the deciding. whether concern- ing the self. return to health. and use of symbols. some six modes of constructing the rational can be demonstrated using well-known field studies. which cannot be understood without . the details become fascinating and self-explanatory. Semiotics The assumption here is that the appearances of the world. or in modern periods. bodily motions and gestures. QUALITATIVE WORK AND CONCEPTIONS OF RATIONALITY Shifting now to the application of conceptions of rationality. The weight of evidence strongly supports an emergent rationality that is grounded in sanctioned deciding that incorporates the mix of values featured in the culture. In this sense. chronic incapaci- tation or death cannot be easily distinguished from the belief system in which they are cast and which always contains an explanation. outcomes and consequences. FRAMING THE RATIONAL IN FIELDWORK 75 of legitimacy. a healing ceremony in south-eastern Mexico among the Zinacantan (Fabrega and Silver 1973) implores the gods to heal or be propitiated. In many respects. In many respects. For example. macabre and isolated cultures. This is background for the modern attempts to explicate in social research the place of rationality. mainly given to us via words. the language games that obtain in pre- literate societies are expressive exercises in which the values are played out in known forms. The expressive and the instrumental aspects of this cannot be disentangled. relatives and healers. the modern state obscures the seething and unresolved conflicts and passions of competing groups. while substantive rationality shapes the ritualistic or broadly ceremonial decisions. gaming or cannibalism. such as canoe building. families. as well as requiring manipulation of religious phrases. Science. an ideal type that must be stretched to cover the instant case or example. has an apparent rationality. sustaining a culture. Weber’s grand synthesis of rationality is tenuous (Sica 1988) because it seeks to identify the governing forces that embed deciding.

but is a technique for describing the logic of symbols. It continues the notion of binary decisions. In this formulation. In many respects. it is displayed at ceremonies honouring people. This connects the semiotics of Eco with the work of Mead (1934). pragmatic deciding is the basis for the links. the ‘code’ or the means by which the messages are seen as com- municating. semiotics is a marvellous tool for simplification and disaggregating of complex materials but over- looks historical variations on usage. Charles Morris. White signifies honouring and reverence. but are built into the analysis. once one discovers. the posited rationality of semiotics found in de Saussure (1966) is the logic of mathematics.76 PUTTING THE PRACTICE INTO THEORY reference to a code. Thomas Sebeok. Consider Mannheim’s (1960) and Garfinkel’s (1967) arguments that rationality was determined by the situation. Context-based rationality Mannheim (1960) and Elster (1983. dead or alive. White stands as a contrast. In Japan. Semiotics is not a method. This argument from situational rationality is consistent with pragmatism in . Variations exist. If in western European societies the colour ‘black’ signifies reverence and a distant thoughtful mood. The continuity in social life arose as deciders look for a pattern at the same time that they seek instances of it in events. the law of the excluded middle. The individual moves are not seen alone. the primitive meaning connections between the signifier and the signified are assumed. 1984. weighing the factors and deciding. his student. or uncovers. The two principled kinds of semiotics differ radically in their assumptions about deciding. reflect back on their previous moves and decisions. pragmatic semiotics. hierarchical and consistent governing principles. First. based on the work of Pierce (1936. economics and Aristotelian thinking. the significa- tion is reversed. The logic of symbols obtains. but these poles are subsumed under the for- mal code as less salient than the mood they convey. 1989. but in a prospective–retrospective perspective. Second. 1958) and more recently of Eco (1979). relies on the identification of the ‘interpretant’ or basis for interpretation that links X and Y. this draws on the Meadian proposition that meaning arises in the completion of the gesture toward an object. signifying a different mood and attitude. and the complex constructions of meaning within given domains. and Morris’s student. not discov- ered. and overemphasizes the constraints of culture’s codes. 1999) argue for a situated rationality. Since the powerful device is the code rather than individual actors’ sense-making activities. These oppositions are based on general readings of culture and have shaped structuralism’s sometimes rather thin caricatures of culture. and project them forward as a pattern to imagine how to play. Given a sign composed of signifier and signified. Playing a video game or chess presents a good example. both connected in the nature of the arbitrary meanings of a given culture. The colours also connote ‘good’ and ‘evil’. as players are both trying to develop a strategy.

Elster explores the forces of rational choice that obtain in spite of irrationality. Because we cannot predict outcomes. and in general obviates questions of power. . In many respects. This latter. greater thought. a logician and historian. Perhaps the most powerful of his writings explores Solomonic Judgments (1989) in which randomness in out- come may be preferred because of limited time or information constraints. which has great promise. Elster. envy and shame. He points out that there are occasions in which non- rational forces shape deciding almost in spite of rational ‘impulses’ which he explores in his Ulysses and the Sirens (1983) or that non-rational motiva- tions lead us to deny things we fancy. of weighing of options will increase satisfaction derived from an outcome. It is the problem of fieldwork. to describe and analyse the social context in which emotions arise. The problem repeatedly is that we are unclear about how rational to be. It is often the irrational attachment to rationality that creates paradoxes. abandoning an attachment to rationality is the best course of action (or non-action). This truncates issues of the competing rationalities that may play into the deciding. a topic of Sour Grapes (1984). Having said that. Short-term and long-term rationality may be in conflict. and so not compete in a commodity race with neighbours. Elster’s brand of rational choice is based on empirical realities rather than hypothetical models or attributed motives and values. requires that our others are also on the whole acting rationally! Elster’s rationality is a situated and contextual rationality which might be called ‘becoming rational’ as situations require modes of self-understanding and ‘predicting’ what the other as a rational being is intending or will do. Jon Elster’s approach explores the patterning of choice by context. We may be required to act irrationally in a market-driven society. that rationality obtains and is reasonable and rational. posits rationality as a kind of working framework. and how they are resolved. intuitive and situated. if we espouse long-run rationality. he concludes that social emotions and their expression are shaped by the social context in which they arise (Elster 1999: 203). decisions are made in line of ‘indirect goods’ or the ‘side bets’ and satisfactions we encumber almost in spite of rational deciding. In this sense. and uses examples from literature. Elster argues that people oscillate from nor- mative guidance via rules to rationality or principles that reflect the ends we wish to accomplish in interaction with others. rare indeed. Conversely. and are wedded ‘irrationally’ to rationality. FRAMING THE RATIONAL IN FIELDWORK 77 general because it locates the dynamics of deciding in the here and now. The time and energy one might devote to an ideal solution is inconsistent with an efficient outcome. random processes may be the closest approximation to ‘just’ or fair decisions. Often. philosophy and history to support his claims. and is independent of discourse. His works explores the conditions under which. there is often in life no guarantee that more informa- tion. jealousy. and the gestures and actions toward the problem. and the consequences of most deci- sions are situated and unfolding. of course. In sophisticated analysis of such key emotions as honour. Much necessary wisdom is grasped. then. Although more structural in orientation.

let us take the idea of ‘negative gratuity’ in Kojeve (Drury 1994) adopted by members of the Collège de Sociologie – Bataille. post-war French thought shaped by Hegel. the rational and rationalizing processes. This position elevates feeling over thought. policing and governance – that resistance is sustained. The being question was often side-stepped in favour of doing. Callois and Leiris. Self-presentation as a driving social force Goffman’s (1959) form of situated rationality is shaped by existentialism. While enlightenment values stressed transcendental knowledge and development. The stripped person . while its themes are quite provocative to explore in the context of organizational life. or non-being. Nietzsche and Heidegger emphasized the consequences of the death of history. The profundity of that proposition shapes the ideas about the nature of society. it is only through opposition to rationality itself. This is mirrored in current preoccupation with risky experience. 1997b [1992]). Its early thrust was to sustain humanity in light of the growing spurious rationality of modern life that suppresses feelings. acting. a reaction to the rationality of both Marxism and fascism arose in France and blossomed in due course into various versions of structuralism (Dosse 1997a [1991]. surrealists such as Breton and Dali and abstract artists such as Kandinsky. is a consequence of dread and fear. and has global and non-empirical echoes of grand phi- losophy. meant that grappling with the meanings of being. and the importance of excess. The underlining proposition. found in Heidegger. and sexuality in particu- lar. presenting an impression and sustaining a self. (These themes are echoed in Foucault. The adaptation to this existential quandary was various: playful and gamelike in the work of writers Georges Perec and Raymond Queneau. Others tried to create a dialogue to assert meaning and significance through an attack on reason and rationality itself. Since rational- ity drives the development of a more market-based society. risk and anti-rationality. The inability to act. a derivative of the above French movements of thought. the non-rational in the sense of the playful or aesthetically meaning- ful. being in the world. Kojeve and their followers. who is the ‘son’ of these luminar- ies of the generation before his own in French universities. the triumph of reason. or non- being. Goffman posited a central dilemma of the actor-being or doing – terms from Sartre’s Being and Nothingness (1953). corrections.) For purposes of illustration.78 PUTTING THE PRACTICE INTO THEORY Higher rationalities Following the Second World War. a dark and romantic idea what might be called the joys of sex. being must be constructed in light of the threat of death. and counter- asserted the negative gratuity. his reworking of Heidegger’s Being and Time (1953). emotions. to assert a self to be validated by others. and a slow developmental evolu- tion. was impossible. in its several forms – education. Stripped of meaning and necessity. was that the denial of death.

Reflexivity reflects weak authority and competing rationalities (Espeland and Stevens 1998) as well as irrational forces at play. long-term goals and purposes blurred. FRAMING THE RATIONAL IN FIELDWORK 79 whose self is at risk for Goffman is a pitiful and pitiable manqué. Giddens poses the reflexive process as the background to sustained rationality. The self is never a thing. In this existential situation. these approaches all remain limited as working fieldwork per- spectives. we see echoes of Heidegger’s notion that man denies his being. Goffman eschews a goals-purpose model of action. This then permits identification of the key deciding points and uncertainties that attend them. Goffman’s rationality is that of the situation. like Elster. . even in organizations. or a loose assemblage of per- formers. or resources. and the strategic being highly bounded. among apparently disparate properties and entities (Espeland and Stevens 1998). obscured by habitus. always becoming. 1991) find rationality embed- ded in process. Notions like agency. and do not trust impressions that can be easily manipulated. because he denies death. using techniques. and rules or means. or a dyad. In summary. or ‘action’. fol- lowing Sartre. This suggests a quest for statistical ordering. Technical solutions to the tensions between rationality and irrationality as system properties have given rise to the current fascination with rational choice theories and the metrification of economists. While self- presentation is a fundamental drive. are invisible ordering devices for these theorists as ‘minds’ and ‘souls’ disappear into the codes of action they circumscribe. yet be unclear and uncaring about the long-term consequences of their actions. 1990. I would suggest an understanding of rationality-in-the-field requires us to first describe in detail the key processes and routines that organize collective action. Goffman. institutional rationalizations and the broader processes of sense-making. In such uncertain circumstances. assumes an ironic pose. Goffman. Revisionist structuralists Bourdieu (1977) and Giddens (1981. Language is a part. There is little room for the negotiating. 1984. caught in temporality and in opposition to the other with whom the self is easily conflated. or the nature of it. it is always contingent upon the vali- dation of others and as such is a compromise formation. scaling and formulae. its dynamics and forms not fully understood. interests. argues for a humanism in which some mutual regard and shared fate is delegated to a team. but only a small part. choice and the evaluative act(s) of deciding prior to examining context and aims. even as a weak alternative to a formally grounded legitimate society. These nevertheless are unsatisfactory because they begin by positing rationality. actor in the scene. and rejects a ‘pipeline theory of communica- tion’ which sorts out messages by their instrumental aim and the rest as ignorance and error. sense-making. rationally. of performing convincingly because people discount what they see. one in which actors can seek expressive fun.

This means that the values and norms that they hold and the working rules that have evolved are employed to pin down meaning. The field. and so on. To some degree. it also must be said that the features of such deciding may in fact be general and generalizable. They encounter a situation ‘primary reality’ and try to see if meanings in context indicate a secondary reality emerging. is located analyti- cally within the surround (Bourdieu 1977). FRAME. specified context. rehearsals and technical . we often have to use cues to make sense of how realities shift. the situation of deciding is a given. In social life. ceremonies. a set of subjective and objective forces. In short. Fur- thermore. FIELD AND SURROUND Frame analysis A useful approach to the investigating rationality in fieldwork is derived from later Goffman (1974). and work ‘down’ to the concept of a frame.80 PUTTING THE PRACTICE INTO THEORY people must decide without adequate or full (or too much) information about the consequences of their actions. Goffman names five frames (drama. grant meaning to objects. play. and to coordinate with one’s colleagues in organizations. All deciding is embedded or layered in social orders. little dramas. energy and the attention of others will not permit it) stock phrases. given a primary face-to-face social reality. this means that schemes of typification and generalization. Often. In this way. In a sense. trust and tacit knowledge. The field is an institutional struc- ture. Given that events are ambiguous. rehearsing them. and constrain dynamics. when deci- sions are complex. The field contains many frames that give rise to action and define the nature of the event or situation confronting the actor. but this remains to be empirically investigated. regrounding of events. or set of norms and procedures that shape careers. and can be cast as stories. the broadest political environment is the sur- round or that which clearly cannot be controlled or altered by those who decide. Only the bare tenets are outlined here. sense making requires a wait and see attitude. clichés and rules of thumb are exchanged to rationalize and account for the deciding done. Goffman argues that all social reality is contextual. I want to use frame in a generic sense. it is not possible to understand rationality outside of a particular. none has a hegemonic grip on ‘reality’. That is. These accounts in turn are only as good as the audience(s) that accepts them. ‘profiling’ are used to reduce uncertainty to manageable dimensions. Beginning at the most distant. as answering the question of what is going on here. organizations are arenas in which situationally justified actions are made non-problematic. since variations in individual decisions can never be fully captured in explanations (time. non-routinized and the technology is ‘human’. one is thrown into life and must come to grips with its meanings. once the features are isolated analytically in that context. Let us begin at the most abstract social level. the surround. actors require meaning. Framing is deciding ‘what’s going on here’. Having said this.

this means that the police are engaged in ‘technical reground- ings’ (Goffman 1974: 58–74) interpreting actions in terms of some new set of motives or purposes not apparent in the ‘primary reality’ as initially framed. whereas when efforts are made to mislead someone the ‘frame’ is a ‘fabrication’. shaped or altered. This process of actually framing a case begins with the first visit to the scene and carries on to the conclusion of the case. the budget processes that define promotion opportunities. My focus here is on explicating the form. class and status. First the surround of policing is the taken-for-granted aspects of society that cannot be changed. arrest and possibly confine people (Fagan and Davies 2000) and these are largely sociological or structural features that constitute the surround within which the fate-shaping stops made by officers unfold. the city’s structure. and more immediately. city politics. or fluctuations in the tax base of a Midwestern university whose lifeblood is the state-allocated budget. the nature of crime. whose meretricious eyes shape actions in advance and generate covering paperwork and accounts. The actor chooses among the choices available. The media. to be discussed below. the composition of the city’s population in terms of ethnicity. . Even if the murder is cleared via confession. media attention. constituted by elements that are defined in or out of the frame. is an attempt to re-create or script dramatically previously enacted events. FRAMING THE RATIONAL IN FIELDWORK 81 regroundings). raises and salaries. Detective work generally as framing Consider further deciding in detective work in framing terms. the framing of the sequence of events is even more essential. a set of assumptions and material constraints defines ‘real policing’ – controlling incidents and producing face-to-face solutions to real time here and now presented problems by means of tactical manoeuvre. and the like. criminals and victims. The surface features of an event as framed or defined against the backdrop of the field and the surround. increasingly constitute it. These are generally consensual in nature. Abundant evidence suggests that police use the ecological and demographic features of cities to shape their decisions to stop. those that are negative realities (those that threaten the tenacity of the frame) and those matters that break the shared frame. think of the budget processes in a large city. If the case becomes a ‘whodunnit’. that results from framing a murder as an investiga- tion to be worked. To some degree. question. the police must frame it as a coherent story or narrative that will enable the case to pass through the criminal justice system. such as the ecology of the city. Police detective work. Homicide framed as a technical regrounding Consider the work of a homicide detective in framing terms. or case-logic. and new hires. Analogously. The policing.

partially intuitive. The victim (a body) must be present. depositions. and interviewed (often a crucial turning point in investigations. if any identi- fied. death pronounced. and are not static ‘occupational culture’ with transituational rules and norms. such as an abandoned car and signs of a struggle. witnesses and suspect(s). The ‘case’ is a concept. a detective ‘catches’ cases as a result of standard organizational procedures – by assignment. once the person is confirmed as dead. The case as a social form The case is the key concept. or a murder-suicide. viewed as highly skilled. As I discuss below. witnesses (no one saw the event. A standard or ‘typical’ homicide case involves a dead body. death has not been pronounced or remains under investigation (e. In this ideal-typical case. suggest ‘foul play’ but no body is found). and loyalty based on rank and organizational position and loyalty based on per- sonal relations. constituted by tradition. The organizing social form is the object of concern within the routines of a detective. interviews. a body is found in an alley several days after death). suspicious circumstances confirmed by the coroner or official. or two people die and it is written off or cleared immediately as a dual homi- cide. a weapon. This ideal arises in some percentage of cases because elements of the case may be missing: the body (people disappear. Subsequent decisions follow from the case-as-focus frame. partially rational and routinized with an aim to ‘clear’ the crime. and by organi- zational processes. These tensions interact in given situations or configurations. itself a social creation. Working for detectives is working a case. a charge may be laid. and binding and blinding perspectives.g. etc. collective obligations and individual ethics and morality. by rotation or special detailing or seconding (as in the UK when a senior member of the Criminal Investigation Department (CID) is seconded to another force to supervise a major case) and works it. a coroner’s jury or hearing is scheduled). or the inter- view and admission used to go to trial. circumstances. shaped by local traditions. In general the policy in detective . evidence (physical and other). are craftwork. suspects. colleagues and the public. Detectives must struggle with a configuration that includes tensions between independence and dependence. they deny knowledge of the murder. missing persons are reported dead.82 PUTTING THE PRACTICE INTO THEORY Second the field for police is the configuration of values and routines that organize their deciding. delayed. ‘Homicide’ defined legally is replaced by working rules that bridge everyday realities and anomalies. and a core mini-drama for detectives). The homicide squad receives its cases. by unspoken and tacit knowledge. witness statements. After the case is worked. an investigation is unfolded: the case is the organizing point by which other materials (evi- dence. declared dead. other evidence (naked body is found floating in a river). Clearance is an organiza- tional definition that is context based (see below). and police notified. Investigative procedures.) are framed and around which an investigation is hung.

support from present detectives. altered the previous dominance of white. FRAMING THE RATIONAL IN FIELDWORK 83 units is that homicides will not be bargained. A homicide unit is composed of plainclothes officers. not by rational deciding. The detective’s ‘case’ is defined by the layered meanings it denotes and connotes in the subjective and objective forces in the surround (the social structure of the city) and field (the occupational values. Detectives investigate. transfers and promotions indicate additional power relations. or ‘politics’. 1993). or considerations based on clearing other crimes. and their salience within the case. The work of the homicide detective illustrates situational rationality. Ericson 1982. how the elements are identified. personalistic and often stylistic. and ‘fit’ with local traditions. When organizations are riven by conflict between ethnic groups. . The content of the frame. more detailed aspects of work. craft-like. following a change in Detroit city politics. or testifying in other crimes. and then focus more closely on the contextual or situational nature of their abiding rationality. usually housed in central headquarters. Anomalies are smoothed by tacit knowledge and organizational conventions. the law. or conventional morality. will not be discussed here (see Sudnow 1965. not the world-at-large. and sensitive to the changing horizon of possibilities that a case represents. process prisoners and present evidence in court. Transfer of black officers into the detective ranks in the 1970s. RATIONALIZATION AND RATIONALITY IN HOMICIDE WORK The organization of homicide work I want to outline the structure and function of police investigative units that deal with homicide and suspected homicide. In fact. Becoming a detective indicates informal power relations in the detective unit. routines and tacit knowledge) and framing or regrounding the events. or case. A detective who is ‘up’ ‘catches’ the assignment (usually on a rotated basis) and proceeds. This form of rationality is case-based. bargaining is routinely done at every point (McConville et al. McConville et al. inter- views and phone calls. a bargain to a lesser charge. Homicide bureaux are considered elite units within police departments. medi- cine. how they are typified. that is the suspect will be given immunity. 1993). catholic Irish investigators. who receive cases primarily from uniform officers or dispatchers who report suspicious deaths or near deaths. good interpersonal networks within the organization. Apprenticeship training as a probation investigator and a brief advanced course once promoted constitute training. is loyal. divisions exist between the top command and detectives. It signifies that the person has the potential to be trusted. and a move into the unit is considered a promotion regardless of rank. The organization clari- fies the nature of the work required. It is a key contingency that alters the horizon of every case that is worked beyond the initial paperwork. Becoming a detective requires personal skill. a ‘team player’ and has investigative skills.

Paperwork is required. to the crime scene to investigate. to ‘produce’ outcomes. Cases assigned to partners are the official workload. In other respects. . or in the normal case.. are processed quickly and given little attention (Audit Commission 1993). It is a powerful reflexive indicator used generally to assess the performance of the unit. an expected volume of paper. The clearance rate for homicide varies across the country: estimates are from 60 per cent to over 90 per cent (Bureau of Justice Statistics 1998). Most reported crimes. size of the unit. Decisions are taken collectively and in consultation with top command by those in charge. . witness statements. and paperwork is demanded at regular intervals (these vary but includes the creation of a case and a follow-up usually within two weeks of the initial investigation). and overtime is distributed by sergeants’ or inspectors’ decisions. containing interviews. The homicide clearance rate itself appears to have little independent association with reported crime rates in general. under rigid time constraints . (Waegel 1981: 263) Detectives are under pressure in general. Very large case files are created and maintained carefully. Working a case A detective’s approach to ‘working a case’ is not fixed. However. like all workers. However.84 PUTTING THE PRACTICE INTO THEORY usually with a partner. . but the order in which tasks are undertaken and the level and direction of effort are left to partners. even while it is routinely manipulated. detectives in medium-sized cities. the grounds of the occupational culture would define how and when to proceed on a case (Corsiaros 1999). . These formal public records are reflected in detectives’ daybooks in which they record their movements and daily activities. control over the case moves up the hierarchy and is controlled by sergeants and above. nor determined by written procedures. . the num- ber of homicides and the percentage ‘cleared’ (an organizational label) vary in urban departments. and is insensitive to the quality of the investigation and court results. their discretion is not unlimited. cases processed under time limits and to cope with a constant input of cases. physical evi- dence. overworked and unable to complete the burdensome paperwork in time. Organizational structure and the detective occupational culture interact over time in case processing. although this varies temporally. individualistic. . if any. The personnel. If a case is defined as politically salient or ‘high profile’. consider themselves swamped. Cases are created and assigned daily. suspect lists and a running record of activi- ties related to the case. unlike homicide. these agents do not generally proceed by following a set of codified rules and procedures. who manage one-half to one- quarter of the workload of detectives in the New York Police Department (NYPD) and Chicago. Waegel (1981) writes: [detectives] typically must process a steady stream of cases . nor are their decisions .

Once categorized as to be investigated. witnesses present. and sergeants and lieutenants seek to increase the clearance rate or the percentage of homicides cleared. but it is context- dependent. and pressures to produce closures if not arrests. practice in individual departments varies as do the warranted processes of producing an organizationally acceptable clearance. as well as the conventional wisdom about the difficulty of the case. The social constraints resulting from the analytic features of the crime and its social organization lead homicide detectives to draw subtle yet parochial distinctions to classify and distin- guish the quality of investigations and the quality of investigators’ work. nor uni- versally applied. I want to discuss aspects of the work that are variable. defines and sets standards for homicides and clearances. In Baltimore (Simon 1991). how creative the detective is in re-creating the crime scene. Performance is indicated traditionally by unit clearance rates. in a collegial and equalitarian organizational environment. failing to clear conventional crimes and adding to others’ workload decrease status. Status accrues as a result accomplishing the possible with style. the workload is monitored and output evaluated. The criteria for clearing a homicide are neither universal. given the context of . Officer and squad keep records of clearances. racial stereotypes also shape prestige. although this is defined contextually and it is unclear what the ‘product’ of a homicide investigation is other than efforts to produce a clear- ance. 1977) suggests that a few key elements – physical evidence. Where ethnicity divides the unit. ducking cases. Unlike most reported crime that becomes a detective’s responsibility. These are rarely under the officer’s control. FRAMING THE RATIONAL IN FIELDWORK 85 Investigation is a priority job. or a confession – are of critical importance in determining the proba- bility of a clearance. Clearing high-profile cases conveys lasting respect. this binary distinction was colourfully called ‘dunkers’ (easy self- cleared cases with witness. While the Bureau of Justice Statistics compiles the Uniform Crime Reports. maintaining evidence files and following hunches. Such mat- ters as ‘flair’. ‘Cleared’ is the central thematic term in homicide units. cases are divided according to working rules in the unit into routine and non-routine. Production norms based on bureaucratic rules about routine and timely filing of reports. Nevertheless. Evaluations do not rely entirely on the percentage of cases cleared. pursuing leads. a reported homicide is not disposed of without investigation. Homicide is not likely to be quickly suspended or abandoned for lack of an immediate likely solution (Waegel 1981: 267). Typification is key to working a case. remain (Wilson 1978). Failure to carry the load. evidence and perhaps even a confession) versus ‘whodunnits’. What. Research (Greenwood et al. Rank or formal position are rarely used to define social relations (Corsiaros 1999). Detectives are expected to ‘produce’. The craft and the work2 Having made a case for the situated rationality of homicide work. is taken into account.

The occu- pational culture assumes that information is virtually personal property and . Knowledge is personal knowledge and often kept secret from others. this is in general the aim of the courts. and reducing the load of others. but no study has detailed the relative skills of individual homicide detectives by any systematic measurement of clearances. officers may wish to be paid time and a half or more depending on the organization. It is perhaps obvious that the skills of homicide detectives vary. Knowledge developed and synthesized on a case is rarely widely shared with others. The number of officers available. if a hearing or trial is scheduled when an officer is on duty no overtime is accrued. or the like. for example by other detective units or uniform division who can lend personnel for investigations or raids. Investigators’ skills vary widely. does the fieldworker want to identify. who can be assigned. those assigned to a case. measure. facts and associations possessed by individual officers that are not written down or shared. while stated in the unit’s budget as fixed. The notion of rotation. and the person who ‘catches’ the case may not be competent. informally or formally. the length of the assignment. major cases solved. nor written and accessible. and shaped by alternative sources of income accessible to the unit from fines or seizures. are actually quite flexible in their use and can be augmented or reduced very quickly by command decisions. transfer pris- oners. intended to equalize the case-workload. Court time and payment for court is defined con- tractually by union rules and is a subtle process of gaming in which the courts. compare and analyse. and are teamed with other more competent officers to reduce the differential in cases cleared (Simon 1991). These officers are expected to close fewer cases and to produce results on a few. Complex cases are assigned to more skilled or senior officers. describe. Workload (cases per officer) is not equal in homicide units because skill in clearing cases is not equal. if it is a complicated case with several suspects. is inconsistent with the assumption that equally skilled officers are assigned to every case. and thus prefer to have it scheduled in their off-duty time (Corsiaros 1999). on the other hand. the police and individual detectives engage. That is. say. The assumption of equal skill is contra- dicted by interviews with homicide detectives and ethnographic evidence. abundant physical evidence and legal subtleties. thus increasing their for- mal workload. Measuring the workload of detectives is a complex matter. either across organizations or variations over time within an organization? I highlight each of the clusters of variables of interest. and the overtime available vary across units and within them.86 PUTTING THE PRACTICE INTO THEORY homicide work. They are also allowed to carry cases longer. In most offices. and to continue to work vexing cases. What might be called collective or institutional memory. Resources are important and rarely assessed. Poorer detectives are likely to have more cases open than others. Resources and their allocation are loosely coupled to the resources available in other departmen- tal units. pay for information. Resources in detective units. cases are assigned in order to a rota of duty officers. increase iniquitous practices.

identify multiple victims of the same mur- derer. and reducing capacity of crime analysts or other officers to infer patterns of co-offending.g. or links between crimes. accessible informant files.J. their middle management and higher command. ‘Working’ a homicide case seems to involve a combination of substantive and concrete local knowledge of offenders. for example. juveniles. The style of an officer is partially determined by workload. Seasoned detectives are more resistant to using expert systems than younger officers use with some computer literacy. drug and/or warrant serving squads. and rarely used physical evidence. associational thinking and some- times pursuing intuitive hunches and following ‘leads’. for example. handcuffs and the law). guns. The working style of detectives. Officers work from the case ‘out’ and develop their cases privately. Justice can be obtained. workload varies the likelihood of certain aspects of an investigation being followed up. bringing satisfaction in occupationally defined terms. These decisions may divide detectives. cell phones. In South Central Los Angeles. The individual detective is seen as the defin- ing ‘expert’ in a case unless it is reassigned. justice and self-righteousness. and their style of working is shaped in part by their values. Motives and satisfactions of detectives vary. and (especially in homicide) the absence of data sharing. as well as fronts and tools (radios. Corsiaros’s (1999) work. Because the moralizing perspective defines the outlines of the case. records and case files of other squads such as gang. homicide detectives had so much work they rarely filled in the paperwork and ran six to nine months behind. Style includes self-presentation and appearance. Simpson case in Los Angeles and the Stephen Lawrence case in London. shows that detectives define the moral character of sus- pects and their own notions of justice. in their view. incompetence. algorithmic and probabilistic thinking. The values and motives of the detectives investigating the case shadowed both investiga- tions. procedures and policies lead to systematic resistance and . as shown in the O. settings. revenge. and tends to increase the pressure to clear the case by means other than an arrest or outstanding warrant. records. for example (Corwin 1996). The interconnection of racism with such moralizing is explosive. meaning how one does things. FRAMING THE RATIONAL IN FIELDWORK 87 secrecy is highly valued. rather than linear. A corollary orientation to computing is unwillingness to enter facts. and types of offences. physical evidence and informants. in the initial brief and cursory interview with Simpson after he was charged) in the other. Detective work is driven by a combination of moralization. Detectives in general enjoy being ‘out of uniform’ and value style in dress. varies. In one case sympathy and identification with the suspect (e. Officers’ approach to cases varies. by failing to act as well as acting. as does their use of com- puters. The central character in that book. thus restricting access to their cases to others. vast cover-up. manner and approach to suspects and witnesses. perjury and failure to identify with the victim and his family (Macpherson 1999). evidence and developments into computer-based case files. changes in laws. Clearly. This is magnified by the absence of articulated databases within departments. preferred to pressure infor- mants to confess as a primary tool.

are not always sought. 1993). murder-suicide or self-defence. the present stated case is merely background to other moral and political issues (Corsiaros 1999: 98). confession or subsequent plea bargain. A well-done investigation may produce better evidence. a closure. and detectives’ aims vary from case to case. Some representative modes of organiza- tional closure include closing by: shifting the case to another jurisdiction or agency (a murder of a service man is shifted to the Federal Bureau of Investigation (FBI) although he was killed in a city). clearing it from organizational records. it might produce good will. or to a lesser charge. These modes for clearing vary across departments. means that the networks of cooperation that ideally homicide detectives use can be quite significant. The rise of victims’ rights groups. deeming it double murder. As this implies. mistakes and errors out of the control of the case officer. repeat victims as well as repeat offenders. While conventional wisdom would elevate the instrumental aim. Even in a matter as serious as murder. many factors out of police control affect the length and effort entailed in an investigation.88 PUTTING THE PRACTICE INTO THEORY avoidance of the foreshadowed consequences. Craft work as well . even given that clearing the case is very salient. on an organization- ally defined outcome: a ‘clearance’. The indices of success are embedded in tacit knowledge about what is ‘good police work’. to merely known to be weak or insubstantial cases from the court or prosecutors’ perspective. The interconnection of crimes in crime-ridden neighbourhoods. or to pro- tect witnesses or victims’ families. It is difficult to assess the efficacy of an investigation or of an investigator. as do the tacit practices producing given closures. including inputting and reporting information on cases. exhausted statute of limitations. For example. ‘effort and style points’ and current expectations in the unit. Closing a case. have brought new attention to the interactional and socio-emotional support offered by police to victims. Homicide investigation is focused primarily. victim- compensation schemes as well as courts allowing victims and families of victims to testify before sentencing is rendered. willingness to provide information to the police on other occasions. but not exclusively. clearances. Clearance is a context-based idea. Perhaps the most obvious fallacious assumption is the omission of factors other than random errors which shape practice. In short. the pres- sure to produce varies by the time of the month and the number of arrests or clearances made thus far in the month (McConville et al. because like all complex human-tasks. issuing a warrant. is but one of several functions of a homicide investigation. The bases for these charges range from framing. Assessment is complex and personalized. charges are laid to discipline people. arrest of a suspect or suspects. Thus. or planting evidence. the quality of the police work involved. and the mutiple victim families. although a stan- dard measure of police detective work. coop- eration in future investigations. confession. there are expressive or socio- emotional aspects of the investigation. like arrests. force them to confess to complic- ity in other crimes. knowledge that another agency has issued a warrant. It depends. many functions are carried out by a homicide investigation. or the wish to hold a case to obtain witness cooperation.

says: You never know what might happen. The clustering of cases in a given time span varies their salience. The definition of a case is very subtle. when. Corsiaros 1999). or to give depositions. emer- gency rooms and with witnesses. and often special task forces are assigned (Simon 1991. hover for days and sometimes weeks before they are an official case. like all officers. The occupational culture of detectives The culture of detectives. autopsies. partially a self-protective pose. It cannot be entirely captured by offi- cial figures based on ‘outcome’ data. especially an abduction or kidnapping and murder of a child (usually female). High-status victims. may not violate statistical norms nor affect the year-long rate. like the occupational culture of policing. Such matters are central to policing because they are focal. and the District Attorney in the USA or the Crown Prosecutor in the UK drops the case. events of social importance but of low statistical probability such as a child kidnapping. yet officers are assigned to the case. or even be interviewed. Some cases that are worked for some time are transferred to another agency such as the FBI. and produce enormous good will if cleared. and require special handling of the media. only to find that the victim lived. and affects the field. The weighting of such cases with respect to work- load and credit is not easily resolved when comparisons across organizations are sought. varies by the size of the organization. thus if a city averages eighteen murders a year. but they will alter public and police concern and activity. A case will be given more time and attention depending on the status. Cases may be worked for some time but witnesses refuse to testify. or the state police. as well as ethnicity . The surround changes. and they are noted as cleared but no credit occurs to a person. raise public anxieties about police effectiveness. for example. FRAMING THE RATIONAL IN FIELDWORK 89 as the police occupational culture expects and rewards coping with the unexpected. but the moral and political meaning of cases is variable. and their appearance at a given time and place is not predictable. cover ups of controversial deci- sions or corruption more generally (Corsiaros 1999). If the victim is a police officer or relative of a police officer. A cluster of murders. gender and age of the victim. Some cases. Police ideology. three in one day may not alter the rate. as I have mentioned. These are ‘red ball’ or ‘big cases’. if they received media attention. it will receive special attention. Routines and algorithms are designed to order and routinize case processing. are especially sensitive to cases that might reveal their ‘dirty laundry’ – crimes involving other officers. may have to spend time in hearings. Detectives. by rank and gender. The differences in the salience of cases are very important. how and to whom. are also the subject to additional police time and attention. a mass murder or a series of rapes of elderly women. The unevenness of the workload creates a sequencing effect. Very young and very old victims are seen as of public concern. or was a suicide or died accidentally.

fade into the background in the actual inves- tigation. Paper representations are salient legal and organizational reality. shape and either hinder or facilitate an investigation. Rank. The core of the work is clerical. feudal and cliquish. manage prisoners (those arrested) and process them. Learning hinges on negative instances. most detective work is clerical. but they have a complex. Investigation is the most salient and dramatically important aspect of the work. and its success. passive or active perjury in court. such as paperwork. While the aim is to close cases. the task generally is to clear a murder case. and avenging disrespect to fellow officers. While emphasis is placed on creative investigation. loyalty to partners. These are givens. horror stories. detective work. For example. risk reduction and even con- cern with repeat offenders and their patterns are of little value since homicide is generally a one-time crime for all but a very small number of offenders. Notions of prevention. routine and boring. but seldom directly ‘overrule’ the decision of a colleague. social structure and the economy. or indirect rather than explicit tutelage. the ecological patterning of the events. 1977). There is a tension between loyalty to fellow officers and the working rules of the occupational culture (Corsiaros 1999: 117ff). skill and social status within the unit do not always cohere. The causes of homicide. Tensions arise in how to assemble personal notebooks to contain consistent stories. not investigative. shifting relationship . and to investigate reported occurrences. those that worked and did not (second-hand renditions or readings) and store these cues as guidelines that might configure their own choices. the routine and emergency are not in balance. detectives learn to work cases by watching other detectives work. In homicide. procedures and ‘politics’ (when higher rank officers decide not to move to indictment after an arrest) is normal because the detective views him or herself as the entrepreneur and expert in the case assigned. constraints within which the investigator must work.90 PUTTING THE PRACTICE INTO THEORY (Martinez. There are dynamics of competition within police units. crime management. While social relations tend to be per- sonalistic. There is an irony in homicide work as in other police work. but the time spent in the office in case processing is far greater. forthcoming). Resistance to new rules. Much of this knowledge and learning is pre-cognitive. Investigative expertise and rank achievement are different. so decisions about the case evolve forward from the incident and take on a new character once an arrest is made. because status accrues to solving big cases. nor are the ‘whodunnits’ common. Paperwork is loathed and it is distrusted. gallows humour and cases marked as ‘how not to do it’. is fundamentally reactive – cases are worked as they arise and with the evidence presented mostly in the first instance (Greenwood et al. making phone calls or organizing evidence. hearing their stories about their cases. sergeants can guide. and not based on rank. and informal rankings of each other are envious and often quite invidious. Occupational culture here means the readings (interpretations) officers give to other officers’ and civilians’ readings. Cooperation and competition subtly are interwoven in practice. even motives.

Command officers in theory deal with ‘policy’. frustration and unfairness to the work. In summary. incompetence and low performance. The competi- tiveness of detectives makes creating linked databases connected to case files resisted. although the features are valued. nor is equal credit given. The perspective is not determinant nor does it contain absolute values. or other illegalities: CYA can produce careful attention to legalities such as obtaining witnesses to support a story. these several features provide the form for working cases. malfeasance. many cases are solved on the spot. Paper written is designed to influence supervisors. attorneys and judges in court and peers. budgeting and analysis. Abstract ideas like justice. fairness and legality are not salient features of the perspective. sergeant and inspector. The resources available vary over time. cases vary in their importance. These are factors out of the control of the detective. checking with supervisors. and officers (here I refer to homicide detectives) know that substantively. Officers in middle management. ‘force mission statements and goals’. roughly in order of the description above. or dismissed after limited investigation. it should be noted. This means all cases are not equal. and great latitude. The case is an object to be framed within a field and worked given a surround. They are woven together in complex fashion in . often in response to citizens’ com- plaints. but they do not encompass the detailed content of ‘solving’ a case. or way of viewing and acting upon the world. or juries. and skills. indictment or formal trials. a fact indicated by shifting resources to an important case for example. because they cannot always hold and are often in conflict in a given case. Think of these con- nections. ‘Covering your ass’ (CYA). attorneys in the prosecutors’ office. What little evidence does exist suggests that the ‘crisis’ style man- agement developed on the street to deal with the immediate incident becomes the paradigm for all kinds of ‘management’ decisions (Chatterton and Hougland 1996). errors. While homicide work is a competitive world. uttering warnings and observing requirements for obtain- ing warrants. they tend to cling to case outcomes and examples. are dealing primarily with mistakes. concrete instances and stories rather than transcendental notions so treasured by philosophers and jurisprudentialists. those ‘above’. The paper reality created or accepted by detectives (as a result of loyalty to a partner for example) governs and is interconnected with the narrative they produce for booking. and observing rules for tap- ing interviews. FRAMING THE RATIONAL IN FIELDWORK 91 to actual and reported events. does not always mean lying. or strategic crime mapping. they have stan- dard forms. but since we know very little about what chiefs and their staffs actually do. suborning to perjury. These several features of the work are loosely associated as an occupational perspective. Their time horizon is set for processing complaints. crime analysis units. it must be assumed that they take a more Weberian approach to long-term planning. experience and style vary. This builds in a sense of cynicism. and frustrates attempts to create expert systems. this is constantly dramatized by shifting cases around and reassigning them.

transitive judgements. in turn. investigations vary and should be compared to other investigative forms. a different kind of rationality is employed when the encounter is written up (if it is). and the reconstructive-imaginative aspect of the work more concentrated and augmented by additional evi- dence. and can be seen only as presenting a family of problems to the officer. time of day. First. colour of the citizens and the officers’ past experiences. as we have seen. detectives work with a case-oriented logic. This kind of deciding takes into account many factors that vary by neighbourhood. but the field is always subject to redefinition. They set priorities and meanings. RATIONALITY IN CONTEXT 3 All this has one central and abiding point for fieldwork: do not trust the written case record without knowing the ‘story’ or context within which it was worked and by whom. Fourth. Officers working in narcotics. as does Espeland (1998). cannot predict and must rely on a set of loosely related tactics by which he or she brings the situation under control and resolves (closes) it. not as collective. and other investigators.92 PUTTING THE PRACTICE INTO THEORY stories that suppress some aspects of an incident and highlight others. Let us first observe. define the boundaries of imagination and the possible. They exist as archaeological fragments within departments. and therefore imagine the shape of the crimes. rooted in different histories. but as semi-visible. grant prestige within the unit (and in often in society). uniformed police use pragmatic. even to experienced investigators. here- and-now logic to decide whether to intervene and how to intervene. the channels through which it is to be communicated. the expected audience for the paperwork. statements and depositions that are gathered after the initial event. in the field or ‘on the ground’. The aims of the encounter vary. Third. developmental patterns and time. that organizations are arenas of ‘contested rationali- ties’. the information they possess once processed and often ambiguous. and the deciding that is defined as rational also varies. The features shift in salience situationally and are interdigitated with the working rules of the job. Mastrofski 1999). They are partially shared with uniformed officers. Homicide investigations present elegant and striking examples of contrasts between kinds of rationalities. and the tacit expectations the officers read off about their work rate and output as well as what laws should be enforced at the current ‘political environment’ or surround (Meehan 2000). internal affairs and special task forces must create their cases. and linear thinking. They are part of the field within which the actual elements of a case are framed. The officer. Their hori- zon or time perspective is somewhat longer. and are in constant tension with commensuration. the social contexts in a big organization vary. consciously articulated practices. but some or characteristically associated with homicide work. it is based on the format in which the information is to be cast. or make . surprise and the unexpected (Bayley and Bittner 1986. Second. in other words.

shifts toward visible conventionally defined results. How this database is used in enforcement or crime analysis is another matter. symbolized by the organization’s struc- ture. Consider again the ostensive purposes of a homicide investigation. if stops are made of a car or person by officers. Current shorthand terms and non-discursive practices (what is done. the argument goes. Setting aside the rather complicated character of formalized justice in the courts. and detectives can note or flag the file if a query comes in (see Detective Smith concerning this . The current practice in these five contexts is not always verbally rational- ized or explained. They problem solve. that it makes an example of punishing murderers. Substantive rationality always shapes the ‘instant’ or present case. or even of producing a clearance. con- cerns and peace of mind. the social world of middle management and higher command. not what is talked about) sustain the working fiction that work always pro- ceeds as usual. and is more sensitive to external audiences. use prospective or longer-term rationality. Formal rationality would define homicide investi- gation as devoted to finding the perpetrator and bringing him or her to ‘justice’. The fiction of uniformity is also sustained by apparently interchangeable databases. it is an exception that sup- ports the rule of common-sense understanding. rationality defined in terms of the values of the organization and of detective works more precisely. Cicourel 1986. in large cities. A society with a high rate of crime can be seen as violent and a source of fear. Fifth. This moral-legal argument connects police work tightly with legality and morality. The police field. Apprentice work relies on oral didacticism. a murder investigation could be said to entail an absolute moral imperative: murder is a fundamental wrong that should be visibly and officially punished. creates homogeneity where heterogeneity exists. and links these two to public values. One might argue that it serves to deter offenders from repeating. it is clear that homi- cide investigation is not exclusively defined by the organization as a matter of arresting and charging a person with a homicide. . In this sense. Arney 1991) to homicide work. and bearing in mind that many social factors other than the evidence that attaches to the charged person shape it. deters potential mur- derers from undertaking or planning a murder. If murder is not punished. and submerges the ongoing conflict and competition between ration- alities. and if an explanation is offered. even with investigative backgrounds. FRAMING THE RATIONAL IN FIELDWORK 93 them happen. This means it is ill suited to the crude expert systems that have been proposed as a ‘solution’ that will make certain more efficient detective work (see Ratledge and Jacoby 1979. . official rules and procedures. For example. or that victims and society are avenged. such as Toronto and Los Angeles. and have far greater flexibility in creating the problem to which they bring their attention. marks societal concern.). Recall that clearance rates for homicide have been declining even as the murder rate (reported homicides) has dropped since 1991 (it is . Homicide investigations are matters of substantive rationality. databases can be accessed from any division or precinct. society is in danger.

only one involves arresting and charging a person with homicide. field and surround. satisfying public feelings of revenge and satisfaction. high politics and ideological trends that enter and leave the field from time to time. These are framed into the deciding by traditional formatting within the organization (its records and categories and classifications). The field in turn is shaped by the largest level of influence or the sur- round. and mollifying politicians demanding an arrest. ‘murder- suicide’ and accidental death. defined as effort and resources devoted to functions other than arresting and bringing to justice perpetrators is well legitimized within homicide units. assignment patterns. the here and now is always shaped and constrained by the shared. Of the many ways of producing a ‘clear- ance’. public and media frustration and anger. skill. those with high-status victims that have press attention and arouse political opinion and top command concern. collective and organizational processes of framing. in part stimulated and created by the media. Furthermore. or may be found not guilty. physical evidence. the sus- pects. the good- will. Each of these modes of opening or closing a case is embedded in rational practices that produce understood and justified outcome data (Cicourel and Kitsuse 1967). especially one which becomes media material such as the Timothy McViegh case.J. The arrested person may well not be charged by the District Attorney. The ways in which the elements of a decision are found within the frame. emphases of particular supervisors (on paperwork. Definitional processes at the crime scene. The frame sits within a field or subjective and objec- tive factors that are organizational in large part. hospital and medical examiner’s offices vary. These are not unpredictable in their arrival. ‘suicide’. Conversely. These are overtime. but they are lurking and their potential is always ambiguous. witnesses.94 PUTTING THE PRACTICE INTO THEORY still double that of 1962!). Once framed. Thus. and to make an arrest in the Jonbenet Ramsey case. or processed further by the criminal justice system. prestige and perhaps public confidence associated with solving a major case. is highly valued and essential to maintaining the police man- date. Simpson of the murder of Nicole Brown Simpson and Ronald Goldman. homicide would be very ineffective. . remains active. a complex situation changes its character. The pattern of ‘irrationality’. closure or teamwork) and pressures to produce generally (as defined within the unit). it would be irrational not to attempt to expand extra resources to solve high-profile cases. Recall also that the modes of clearing cases vary within and across organizations. and so on. Other matters are thus ‘framed out’ or excluded from the decision. given the potential police increments of goodwill. and with avoiding public scandal or embarrassment as a result of their errors. Detectives link self-esteem and aesthetics (flair and quality inves- tigations) with solutions of major cases. the set of forces such as the media. if one were to use cost-effective measures to assess the pro- duction of clearances. To push this further. but it tops the table within police departments for effectiveness. producing different meanings of ‘homicide’. Consider a homi- cide investigation with a set of attendant facts concerning the victim. associated with the failure to convict O. As high-profile cases show.

efficaciousness and exter- nal standards imposed on workers. set in a surround. field and framed as ‘cases’. and increased surveillance and control of uniformed officers. On the other hand. It manifests many rationally articulated processes such as the processing of prisoners for court. and sensitive to a tightly knit. provide the phenomenological structure within which what non-police call (often incorrectly) ‘murder’ is seen. rationality. turning points and mistakes that are possible at each stage in deciding. Record-keeping systems for arrests are increasingly computerized and widely accessible to officers. FRAMING THE RATIONAL IN FIELDWORK 95 CONCLUSIONS In many salient respects. sensitive to surround. laying of formal charges. and the social organization of reactive dispatching. • Distrust arguments from efficiency. Their case- oriented. isolated. • Recognize that not everything can be explained. especially homicide work. It is the task of the investigator to illuminate this context comparatively. glossed as ‘context’. substantively rational work must also meet legal rules and proce- dures. It also covers a pattern of choice that is apparently rational. • Recall that making sense of a case is not the same as making sense of cases. • Ask how things are done. • Discover the field within which they are given meaning: what are the subjective and objective constraints on working a case? • Focus on what is out of sight: context. The situational rationality that obtains. The following then become working rules in the field: • Find how cases are defined and framed. • Always ask for the story ‘behind’ a case. people work. organi- zed around clearing cases. maintaining secrecy about decisions. policing is a pre-modern occupation with sacred aspects because of its links to the law. face-to-face group of colleagues who are their teammates. • Follow through a handful of cases to identify the key contingencies. morality and the state. the unwritten and the ‘obvious’. These features of the work. simplified complexity and tightened links. operating on a case basis. a covering rationality that enables them to dispose of their cases. is functionally suited to subtle tasks. . field and frame as perspective. detective work. assumptions. but subtly irrational in operation. verbalized and neatly ordered. These formal and collective rationalizations encapsulate investigative work. that is the exchange of information between units within departments. In these processes. working with partners in loose supervision. and driven by a mixture of rank authority. They shape the operating context for transactions with uniformed officers who produce cases for all investigators except vice and narcotics. informa- tion technology has reduced time. and readings of trust. the unspoken. not why. maintains loose coupling with other units. colleagueship and feudal loyalties.

Arney. Deeper problems arise when such systems are expected to function flawlessly. Barley and Orr 1997). I am presently involved in a lengthy study of crime mapping and crime analysis in a large eastern city. is being forced on the police in the UK. computer designing of structures. These studies ironically describe technologically shaped work such as xerox repair. Bourdieu. computer programming and cat-scan operators. 2 A series of contemporary ethnographies identify the paradoxes created by ‘ration- alizing’ craftwork (Orr 1996. market efficiency and the commodifica- tion of public services. Brandl. Corwin 1996). E. unpublished PhD dissertation. NOTES 1 In what follows. and recent unpublished fieldwork on the social organization of detec- tive work (Brandl 1989. is shaping many forces in Canada and the UK. Barley.96 PUTTING THE PRACTICE INTO THEORY • Learn how to work a case yourself. . (1991) Experts in the Age of Systems. W. (1986) ‘The tactical choices of police patrol officers’. pressures of the market and market rhetoric. Michigan State University. I draw on fieldwork (Manning 1979. Cambridge: Cambridge University Press. D. and Orr. Greenwood et al. Bayley. Settings. suggests that information technology can increase productivity. accounting and budget- ing. Martinez 1997. Rationalization is a currently powerful force because it is connected tightly to the modern ideology of neo-liberalism.S. 1977. and Bittner. (1989) ‘Detective work’. Simon 1991. Waegel 1981. NY: IRL Press. implementing career planning assistance to young officers. 2001). S. not one. REFERENCES Antonio. save time and ensure systematic record keeping. Ericson 1982. linking considered objectives with indicators. and by governmental and executive pressures from the Home Office in the UK and the Solicitor General’s office in Canada. In particular. The combination of pragmatism and using market records to measure the production of ‘products’. Australia and Canada by budget constraints. and avoid errors that might endanger lives. (1995) ‘Nietzsche’s anti-sociology’. Albuquerque. American Journal of Sociology 101: 1–43. taking into account the role of crime and fear of crime in shaping public confidence. police are struggling with rationalization in the guise of setting priorities and more precisely allocating the call load. NM: University of New Mexico Press. S. smoothly integrate operator and machine. (1977) Outline of a Theory of Practice. J. when used in connection with public services. Corsiaros 1999). • Many stories float around an organization. R. employ- ing management by means of stated written objectives. Information-driven policing. Ithaca. College of Social Science. 3 Rationalization. and establishing crime-management initiatives. Journal of Criminal Justice 14: 329–348. (eds) (1997) Between Craft and Science: Technical Work in U. P. studies of detective work (Skolnick 1967.

CO: Westview. M. Toronto. Giddens. FRAMING THE RATIONAL IN FIELDWORK 97 Bureau of Justice Statistics (1998) Homicide Report. unpublished report. Department of Sociology. Bloomington. 2. (1998) The Evolution of Sickness and Healing. C. vol. Lexington. MN: University of Minnesota Press. Stanford. and Leng. A. Garden City. Glassman. W. (1990) The Consequences of Modernity. Princeton. Cicourel. and Stevens. F. Eco. London: Hutchinson. London: Routledge and Kegan Paul. unpublished PhD. J. Elster. P. (1967) Suicide. NJ: Prentice-Hall. (1994) Kojeve. U. Elster. Garfinkel. Elster. New York: Bantam. (1999) Alchemies of the Mind. Heidegger. (1963 [1953]) Being and Time. A. CA: Stanford University Press. (1967) Studies in Ethnomethodology. R. . Corsiaros. and Petersilia. NY: SUNY Press. Annual Review of Sociology 24: 313–343. (1959) The Presentation of Self in Everyday Life. New York: Basic Books. J. Englewood Cliffs. University of Manchester. and Hougland. Lemert. (1967) ‘A note of the uses of official statistics’. M. Greenwood. J. New York: Basic Books. 1945–1966. Dosse. (1973) Illness and Curing in Zinacantan. Cambridge: Cambridge University Press. Goffman. NY: Doubleday Anchor. Schweder (ed. translated by D. Chicago: University of Chicago Press. Dosse. Espeland. Corwin. Goffman. (1998) ‘Commensuration as a social process’. MN: University of Minnesota Press. J. E. H. (1984) Sour Grapes. J. Sanders.. Douglas. McConville. (1984) The Constitution of Society. P. Glassman. R.’. CA: Stanford University Press. Cambridge: Cambridge University Press. Minneapolis. Berkeley. (1974) Frame Analysis. IN: University of Indiana Press. York University. R. Ericson. Fagan.C. M. Geertz. Albany. DC: US Government Printing Office. (1977) The Criminal Investigation Process. Berkeley. (1973) The Interpretation of Cultures. Drury. J. and Davies.. S. A. Fabrega. vol. W. (1981) The Class Structure of Advanced Societies. Minneapolis. (1998) The Struggle for Water. 1. A. F. D. Chatterton. Giddens. Toronto: University of Toronto Press. Boulder. A. C. (1997a [1991]) The History of Structuralism: The Rising Sign. H. Washington. (1995) Sociology after the Crisis. Cicourel. J. (1993) The Case for the Prosecution. J. A. Fordham Urban Law Review 28(4): 457–504. (1999) ‘Detective work and high profile cases’. (2000) ‘Street stops and broken windows’. MA: D. (1996) Homicide in LAPD. (1983) Ulysses and the Sirens. translated by D. . J. Henry Fielding Centre. Chaiken. Palo Alto. A. . CA: University of California Press. (1979) The Theory of Semiotics. Fabrega. Heath. M. Cambridge: Cambridge University Press. and Silver. CA: University of California Press. (1991) Modernity and Self-Identity. (1989) Solomonic Judgements. Espeland. Cambridge: Cambridge University Press. Chicago: University of Chicago Press. (1996) ‘Strategic management’.W. Elster.M. and Kitsuse. H. Giddens. Cambridge: Polity. (1986) ‘Expert systems . E.) Cultural Systems. NJ: Princeton University Press. in R. New York: St Martin’s Press. (1982) Making Crime. Giddens. C. Social Problems 11(Fall): 132–139. M. (1997b [1992]) The History of Structuralism.

London: Routledge and Kegan Paul. Cambridge. P. Sociological Quarterly 37: 101–118. (1968) Dying. (1992a) ‘Technological dramas and the police’. Brown. S. D. F. Waegel. MA: Harvard University Press. Boulder. R. Mead. vols I–VI.98 PUTTING THE PRACTICE INTO THEORY Macpherson. and Jacoby. Sir William (1999) Report on the Stephen Lawrence Inquiry. D. A. (1979) Artificial Intelligence and Expert Systems. T. Manning. G. K. Irrationality and Social Order. CT: Greenwood Press. Cambridge. Mastrofski. J. C. Sechehaye. Englewood Cliffs. Ratledge.Q. in M. R. (1934) Mind. Cm 4262. E. (1988) Weber. Chicago: University of Chicago Press. Boston. Sociological Quarterly 39: 351–368. Morris (eds) Modern Policing. (1979) Narcs’ Game. J. Social Problems 12: 255–276. Pierce. Berkeley. R. . (1999) Policing for People. MA: MIT Press. Skolnick.K.K. (1953) Being and Nothingness. Sica. (1991) Homicide. MA: Harvard University Press. Ithaca. Tonry and N.K. Cambridge.H. W. (1960 [1936]) The Structure of Social Action. New York: Philosophical Library. Ritzer. NY: Cornell LIR Press. P. P. Sudnow. Chicago: University of Chicago Press. New York: John Wiley. Westport. Social Problems 28: 263–275. CO: Westview. Manning. vols VII–VII. CA: Stanford University Press. (1960) Essays in the Sociology of Knowledge. (2000) ‘The organizational career of gang statistics’. politics and the axial media event’. Merelman. NJ: Waveland. London: Stationery Office. Simon. G. (1965) ‘Normal crimes’. translated by C. Meehan. (forthcoming) Homicide Work in Six Cities (working title). C. Prospect Heights. Wilson. Sartre. Martinez. Stanford. NJ: Prentice-Hall.K. Poster. (1999) Enchanting a Disenchanted World. MA: Little. P. P. Pierce. New York: Basic Books. J. (1990) The Mode of Information. (1992b) ‘Information technology and the police’. (1996) Talking about Machines. (1996) ‘Dramaturgy. Orr. Sudnow. Thousand Oaks. Sociological Quarterly 41: 337–370. (1967) Justice without Trial. Martinez. (1958) Collected Works. DC: Police Foundation. (1997 [1977]) Police Work.J. J. J.K. (1978) The Investigators. New York: Free Press. (1981) ‘Case routinization in investigative police work’. Manning. CA: Pine Forge Press. Self and Society. Bally and A. New York: McGraw-Hill. (1936) Collected Works. Parsons. Criminology 30: 327–345. D. (1998) ‘On legitimalaise in the United States’. Manning. Manning. Mannheim. Saussure. Washington. CA: University of California Press. de (1966) Cours. M. (1997) Unpublished field notes on Miami-Dade police homicide detectives.P. A.

Interaction provides the foundation to social organization. the details of its production remain surprisingly disregarded. and it is by virtue of social inter- action that clients receive. For Hughes. It is extraordinary to consider that Goffman in his presidential address to the American Sociological Association in 1982. and perceive. his colleagues and stu- dents powerfully demonstrate. Weber and Parsons. strategies. a substantial body of naturalistic studies of work and occupations emerged which began to chart the routines and realities involved in everyday organizational life.4 ANALYSING INTERACTION: Video. Simmel goes so far as to suggest that the ‘description of the forms of this interaction is the task of the science of society in its strictest and most essential sense’ (Simmel 1950: 21–22). It is through social interaction that organizations emerge and are sustained. The commitment of Hughes to interaction as ‘the subject matter of socio- logy’ reflects a long-standing recognition within the discipline of the impor- tance of social interaction to human existence and sociality. It pervades the writings of major figures such as Durkheim. more than a century after the . it is a consequence of social interaction that people develop routines. Hughes. Despite the importance that sociology ascribes to social interaction. through numerous empirical studies. it underpins substantive contributions across a variety of fields and yet largely fails to form a topic of inquiry in its own right. Hughes. Due in no small way to the lectures and essays of E. The main business of sociology is to gain systematic knowledge of social rhetoric’ (Hughes 1971: 508). Conversation of verbal and other gestures is an almost constant activity of human beings.C. ethnography and situated conduct Christian Heath and Jon Hindmarsh Some of the finest ethnographic studies of organizations emerged in Chicago following the Second World War. the ways in which work is thoroughly dependent upon and inseparable from a tacit and emergent culture which is fashioned and refashioned in the light of the problems that people face in the routine accomplishment of their day-to- day work. it informs socio- logical theorizing. social interaction lies at the heart of organizational life. For example at one point Hughes suggests: ‘The subject matter of sociology is interaction. goods and services in ways defined by the organizations and its occupation(s). practices and procedures.

tools. Becker 1963. bodily and of course mate- rial resources. We would also like to mention in passing the relation between more conven- tional ethnography and studies of ‘talk-in-interaction’. Roth’s (1963) treatise on the treatment negotiation in tuber- culosis clinics.g. Consider for example the powerful study by Strauss (1964) concerned with the organization of psy- chiatric care. Atkinson 1995). We have chosen this example since it is a domain which has been subject to a substantial body of ethnographic or qualitative research from a range of standpoints. felt it necessary to plea for the study of ‘the interaction order’ (Goffman 1983). technologies and the like.100 PUTTING THE PRACTICE INTO THEORY emergence of a discipline of sociology. Glaser and Strauss 1965) have had a profound influ- ence on the ethnographic research. This chapter discusses why and how we might collect video recordings of everyday activities for research purposes and proposes a number of ana- lytic considerations or assumptions which might help inform the analysis of relatively fine details of social interaction. texts. It is also a setting which is familiar to us all and does not require a lengthy introduction or explanation. In this chapter. and explore some ways in which we might interweave one or two concerns within these very different approaches. we consider the ways in which the social and interactional organization of everyday activities can be subject to detailed scrutiny. ETHNOGRAPHY AND INSTITUTIONAL TALK The professions in medicine and medical practice formed topics of particu- lar interest to Hughes. It is concerned with how we can use video recordings of everyday settings.g. such as objects. It is not however solely or even primarily concerned with the analysis of talk. and Goffman’s (1961) influential analysis of the career of mental patients. To begin however it is perhaps helpful to discuss why video-based field studies may provide a distinctive contribution to our understanding of the medical consultation as well as of course a broad range of other everyday activities. He suggests that the aims are . his colleagues and students. their analytic con- cerns and their conceptual distinctions pervade more contemporary studies of medical practice and occupational life (see e. but rather with ways in which the production and interpretation of action relies upon a variety of resources – spoken. In characterizing his own work and the studies of his colleagues and students. and their approach. to explore the ways in which participants accomplish practical activities in and through interaction with others. These and an extraordinarily rich array of related studies (see e. The example we will discuss is drawn from a medical consultation in general practice. Strong 1979. It then focuses on the exam- ple. The discussion draws from ethnomethodology and conversation analysis. aug- mented by more conventional fieldwork. Hughes neatly summarizes their concerns and approach to the analysis of work and interaction. Davis 1963.

So for example in medicine. and their concern with social interaction. ten Have 1991. Bergmann 1992. diagnosis. Pilnick 1998). In general. and there is a growing body of studies of talk and interaction in such areas as oncology. research in conversation analysis has been increasingly concerned with ‘institutional interaction’ and in particular ‘talk at work’ (see for example Boden and Zimmerman 1991. The issue is not simply one of detail. Medicine. treatment and advice are accomplished in and through interaction. and chart the practices and procedures. let alone a way forward for studies of work and interaction.g. conventions and reasoning through which patients and practitioners produce and make sense of their everyday practical activities in concert with each other. ‘career’. Peräkylä 1998. or as computer scientists sometimes say ‘granularity’. gloss the very phenomena that they are designed to reveal. but rather that the emergent. Drew and Heritage 1992). ‘shared understanding’. (Hughes 1971: 240) Despite their commitment to social interaction and the organization of everyday practice. or at least studies in the genre of Hughes and related research. and the bargains which were made – consciously or unconsciously – among a group of workers and between them and other kinds of people in the drama of their work. These studies powerfully reveal the ways in which a broad range of activities such as investigation. the sanctions which they have or would like to have at their disposal. even ‘interpretative framework’. Social interaction is placed at the heart of the analytic agenda and yet the very concepts which pervade certain forms of ethno- graphic research. concepts such as ‘negotiation’. There is not the space here to discuss these issues in any detail but it is per- haps useful to mention one or two points that bear upon the following dis- cussion. instruction and coun- selling (e. medical practice and the delivery of health care has become a particular focus of these studies. Emerging through ethnomethodology. psychiatry. it is not at all apparent that for ethnographers such studies provide a satisfactory contribution. whether it is doctors and patients within the consultation or the members of different pro- fessions engaged in highly technical activities. ‘trajectory’. nursing. ANALYSING INTERACTION 101 to discover patterns of interaction and mechanisms of control. the tension derives from the very different idea ‘context’ . fails to get to grips with the practical and concerted accomplishment of work – that is to examine and explicate the interactional and contingent character of practice and action. Maynard 1992. Heath 1986. practical and contingent accomplishment of work and occu- pational life disappears from view – from analytic consideration – in these fine ethnographies. the things over which people in a line of work seek to gain control. however. in particular talk. general practice. Despite the analytic richness of these studies. there has been a growing recognition that ethnography. ‘bargaining’. Over the past few decades we have witnessed the emergence of a very dif- ferent body of sociological research – a corpus of studies which have attempted to examine in fine detail the social and interactional organization of everyday activities. professional practice is accom- plished at least in part through people talking with each other.

for example. and in other ways articulate an action or produce an activity. performance. but it is not clear that it provides the resources necessary for the analysis of social interaction. feature in the practical accom- plishment of social activities. point to objects. but undemonstrable. Fieldwork is critical to any research which is concerned with investigating specialized forms of social activity and settings with which researchers may be unfamil- iar. and the physical setting appears to produce a denuded characterization of conduct. qualitative sociology has drawn upon field studies and in particular (non)participant-observation of naturally occurring activities in everyday settings. and if it is relevant to consider how people orient bodily. In particular. We also wish to point to the critical import of undertaking field- work as well as collecting recordings when undertaking studies of special- ized forms of activity such as medical practice. In this chapter we wish to address one or two of these tensions by dis- cussing the ways in which we can begin to consider how bodily conduct and material features of the setting. Moreover. ANALYTIC CONSIDERATIONS Since its inception. For conversation analysis. we explore the ways in which talk is inextricably embedded in the material environment and the bodily con- duct of the participants. negotiation. as well as talk. the organi- zation of conduct. grasp artefacts.102 PUTTING THE PRACTICE INTO THEORY found within these bodies of sociological research and ways in which particular characteristics can be said to play upon. participants themselves are unaware of the ways in which they organize their conduct in interaction. or feature in. it is not possible to recover the details of talk through field observation alone. In this way we hope to illustrate how characteristics of the setting which are given some importance in many ethnographic studies may be reconsidered or respecified using a rather different analytic framework. labelling.g. and how objects and artefacts such as paper and pens become momentarily relevant within the course of particular actions and activities. the seemingly narrow focus on talk. role distance and the like (see e. a rigorous commitment to demonstrating empirically the relevance of particular features of the context to the actual production of action by participants in interaction removes any liberal appeal to an array of potentially. the wider organizational framework. they are inevitably engaged in the topic or business at hand rather than the analysis of the ways in which it is . to a large extent. For ethnography. ‘broader’ contextual characteristics (see Silverman 1999). For example. Fieldwork has provided the critical resource for the dis- covery of indigenous social organization and a whole assembly of concepts and theories have emerged which are richly suited to characterizing obser- vations – concepts such as career. it is unlikely that one can grasp little more than a passing sense of what happened. Hughes 1958. and disregard of such potentially relevant features as the iden- tities of the participants and their professional background. Goffman 1967).

ANALYSING INTERACTION 103 being accomplished. Unlike more conventional ethnographic data. Video recordings help provide those resources. specific approaches. the competent accomplishment of social actions and activities is dependent upon participants glossing the very ways in which they produce and recognize conduct. additions to records. We can for example see people writing documents. suggests that we need additional resources if we hope to explicate the details of human conduct in its ‘naturally occurring’ environments. and to recover the ways in which they orient to and handle objects and artefacts. as Garfinkel (1967) and in a rather different way Goffman (1963) demonstrate. video recordings provide researchers with the opportu- nity to analyse the emerging characteristics of those ecologies. and the like. as it emerges within its ordinary ecologies. They allow us for example to track the emergence of gesture. manipulating objects. they are not simply tied to particular projects. Thus. the use of recorded data serves as a control on the limitations and fallibili- ties of intuition and recollection. They allow us to capture versions of conduct and interaction in everyday settings and subject them to repeated scrutiny using slow motion facilities and the like. computers and fax machines. pointing to a picture. it is still relatively rare to find video-based field studies within qualitative sociology. modifications to plans. For those with an interest in the material settings in which action and interaction arises. to determine where people are looking and what they are looking at. it exposes the observer to a wide range of inter- actional materials and circumstances and also provides some guarantee that analytic considerations will not arise as artefacts of intuitive idiosyncrasy. selective attention or recollection. The tacit. In social anthropology there has been a long-standing interest in using first film and . video recordings can provide the opportunity of developing a database which can be subject to a broad range of analytic concerns and interests. coupled with the complexity of action and interaction. a problem which has long haunted more conventional ethnographic research. using arte- facts such as telephones. even though such actions are dependent upon an array of social and interactional competencies. (Heritage and Atkinson 1984: 4) Despite the potential opportunities afforded by video. They also pro- vide the opportunity to show the data on which observations are based to other researchers and subject their analysis to the scrutiny by members of the academic community. In sum. they pro- vide access to the fine details of conduct. stepping on an escalator. both talk and bodily comportment. So for example consider the way in which we can provide only the most cursory insights into the pro- duction of the most seemingly banal activities such as asking a question. or the interests of a particular researcher. Indeed. or experimental design. ‘seen but unnoticed’ character of human activity and social organization. Video recordings therefore pro- vide us with a resource with which to analyse ‘situated’ action. we can also recover changes on screens such as computer or television monitors.

including symbolic interactionism. but rather a methodological orientation from which to view natu- rally occurring activities and events. how- ever. Second. a set of clearly formulated techniques and procedures.g. In anthropology. such as visual alignment. Heath and Luff 2000. Hindmarsh and Heath 2000). First. refer to and invoke objects in interaction. talk is inextricably embedded in the material environment and the bodily conduct of participants. This curious absence of video as an analytic resource derives perhaps more from the absence of a relevant methodological orientation than a lack of interest in exploiting sociological possibilities of video. In face-to-face interaction. Before illustrating the approach. in short ‘methodological resources’. talk and bodily conduct are social action and are the primary vehi- cles through which people accomplish social activities and events. the conceptual and ana- lytic commitments found within a substantial body of qualitative research. it is perhaps helpful to provide a brief overview of three of the key analytic orientations found within ethnomethodology and conversation analysis (Garfinkel 1967. and the immediate environment provides resources for making sense of the actions of others (see e. in many cases. Goodwin 1981. which richly illuminate materials generated through conventional fieldwork. like sociology. Heath 1986). Like other forms of qualitative social science they do not involve a method per se. Nevertheless.104 PUTTING THE PRACTICE INTO THEORY more recently video as a way of presenting activities and rituals. the sense and significance of social actions and activities are inseparable from the immediate context. Indeed. they emerge moment by moment reflexively creating the context in which they arise. Participants point. In face-to-face inter- action therefore. the resources through which we can begin to exploit video for the purposes of sociological inquiry are provided by ethnomethodology and conversation analysis. activity theory and so on. artefacts and technologies. of the co-participant(s) (see e. Heath and Hindmarsh 2000. . social actions and activities are accomplished through a variety of means. gesture often works with and within particular utterances to accomplish a particular action. through which they produce social actions and make sense of the actions of others. there is not a significant tradition in using either film or video as a vehi- cle to actually analyse the organization of social action and interaction. visual and tactile. spoken. Third. Schegloff and Sacks (1974) argue that the concern with talk in conversation analysis does not derive from an interest in language per se. do not neces- sarily resonate with the details of activities one confronts with video record- ings of everyday settings. bodily conduct and the material environment plays a critical part in the production and intelligibility of social action.g. but from the recognition that social actions and activities are accomplished in and through talk-in-interaction. and the documentary programme now plays an important part in both academic and popular studies of ‘other’ cultures. participants use and rely upon practices. and turns at talk are delicately coordinated with the visual conduct. Sacks 1992). procedures and reasoning. So for example. they use tools.

they treat context as the product of the participants’ actions and activities. ANALYSING INTERACTION 105 Unfortunately. noticed and the like. to treat context as a framework in which action takes place. Participants constitute circumstances and situations. This interest reflects a long-standing concern in the social sciences with context and the uniqueness of events and activities. In interaction participants produce their actions with regard to the con- duct of others. their action forms the framework to which subsequent action is oriented. invoked. For example. which are seen. that the objective reality of social facts as an ongoing accom- plishment of the concerted activities of daily life. it is somewhat surprising that in the social and cognitive sciences. is for members doing sociology. ‘in and through’ their social actions and activities. an individual’s shift in orienta- tion may be sensible by virtue of the ways in which it is aligned towards an object such as a picture. It also. Ethnomethodology and conversation analysis adopt a rather different approach. (Garfinkel 1967: vii) With regard to context and situation it is critical to note the concern with the ‘ongoing accomplishment of the activities of daily life’. is dependent upon the course of action in which they become relevant. in recent years there has been a growing interest in what has become generally characterized as ‘situated action’. even among more radical analytic developments. Rather than treating a particular situation as a framework in which conduct takes place. and the ways in which participants invoke and rely upon ‘physical’ features of the ecology to produce actions and make sense of each other’s conduct. including the physical environment and purpose of the occasion. however. Features of particular context. Garfinkel (1967) suggests for example: in contrast to certain versions of Durkheim that teach that the objective reality of social facts is sociology’s fundamental principle. So for example participants produce actions with regard to the . in turn their actions and activities in part reproduce the characteristics associated with particular situations or contexts. The actual significance or meaning of these objects and artefacts. Unlike other forms of qualitative inquiry. like much sociological research. disregards the immediate environment. the lesson is taken instead and used as a study policy. In turn. are thought to bear upon the organization of the participants’ conduct. So. ethnomethodology and conversation analysis are con- cerned with the ways in which social actions and activities emerge moment by moment. Indeed. used. research on communication and interaction. with the ordinary artful ways of that accomplishment being by members known. a fundamental phenomenon. research on non-verbal communication tends to separate conduct into different channels and to some extent disregard the ways in which talk and bodily conduct are interdependent in the practical accom- plishment of social action. There is a tendency. has largely disregarded the ways in which the immediate ecology features in the accom- plishment of social actions and activities. and taken for granted. situations and circumstances are ongoingly accomplished by participants themselves from ‘within’ those settings. activities and events. and in particular the immediately preceding action or activ- ity.

analysis is directed towards explicating the resources.. engender. They still walk round like they know although they are walking around in a dream world. Within the unfolding course of sequences of interaction then. people think they know but they don’t. and is designed to build possibilities for conduct. You find that they all begin by saying something like this: ‘About what I am going to talk about.. the methods. it is understood by virtue of its location with regard to preceding action(s). they are ‘noticeably’ or account- ably absent. and how an answer is recognizable and acceptable by virtue of its juxtaposition with a question. Interaction. Actions are also prospectively oriented. or how. Heritage (1984) refers to the ‘context sensitive–context renewing’ character of action-in-interaction and points to the emergent. participants build an ‘architecture of intersubjectivity’ (Heritage 1984: 254) in which they display their ongoing orientations towards the business at hand and the emerging turns at talk. actions which are relevant within particular locations. as in the example discussed later. what is it that people seem to know and use? Here what people know and use is not to be mapped for each area onto what it is that science turns out to know. The sequential organization of action in interaction is a pervasive feature of the ways in which partici- pants both produce and make sense of each other’s conduct. verbal quote in Hill and Crittenden 1968: 13) Through detailed scrutiny of particular cases. More generally. What we are interested in is. through which participants produce their own actions and make sense of the actions of others. The analytic focus of these investigations is with the practices and reasoning. (Sacks. close inspection of a patient’s eye is sensible and legitimate by virtue of the patient pointing to the object and describing the difficulty. if they do not occur. provides unique opportunities . and therefore to the analysis of social action and activity. the competen- cies.106 PUTTING THE PRACTICE INTO THEORY prior action and the frame of relevancies that it establishes. even elicit subsequent action. but is to be investigated itself. the emergent and sequential character of conduct. Furthermore if you tell them it doesn’t change anything. and which in some cases. upon which people rely in participating in interaction. flexible and contingent organization of conduct. The sequential location of action within the emerging course of interaction is critical to the production and intelligibility of conduct. moreover.’ . Consider for example the ways in which questions establish the sequential relevance of an answer. As an approach. Particular actions establish the sequential import of specific actions by co-participants. designed to encourage. fragments of action and inter- action. which in turn form the basis to the partici- pants’ assessment of each others’ and their own conduct. Sacks suggests that this stands in stark contrast to the majority of the social scientific endeavour: A curious fact becomes apparent if you look at the first paragraph – it may occur in the third paragraph – of reportedly revolutionary treatises back to the pre- Socratics and extending up to at least Freud. action within interaction provides opportunities for subsequent action.

and the sorts of tools and techno- logies they rely upon to do their work. the ways in which they deal with the actions of others and co-participants respond to their own actions. to become familiar with the setting it is often necessary to undertake extensive participant and non- participant-observation. ANALYSING INTERACTION 107 to explicate these resources. and in many cases to have lengthy discussions with participants themselves. and ordi- narily undertake successive periods of fieldwork and recording over some period of time. It is necessary to undertake small amounts of fieldwork prior to recording in order to be able to decide where to place the camera and microphone so that the most relevant (views on) activities are . In this way. includes developing an understanding of the technical medical jargon that is used and how it is deployed. So for example documents. and the purposes to which they are put. the events with which they deal. Interaction provides us with the resources with which to begin to systematically examine the participants’ relevancies. Zimmerman and Pollner 1971. 1974). the field researcher may find it difficult to understand a range of potentially relevant activities that feature on the recordings. Sacks et al. the sequential organization of interaction is both a topic and resource in these investigations (see e. the categories of information they provide. Understanding the events and activities in medical settings. and which may be subject to elaboration. We can examine subsequent action(s) and activities to examine how participants themselves are responding to each other’s con- duct. Therefore. such as a patient’s medical record. we conduct fieldwork before recording. clarification or repair.g. for exam- ple. manuals and log books are a feature of many organizational environments and play a critical part in the ways in which participants organize and report activities and events. For example. In our own research. and in turn how participants respond to the responses of others. research on interaction in complex organiza- tional environments requires the researcher to become familiar with the setting. In many settings therefore it is critical that video recording is coupled with extensive fieldwork in which the researcher becomes increasingly familiar with the characteristics of the environment unavailable through record- ings alone. an understanding which is oriented to in subsequent action. records. It is necessary to understand the sorts of activities in which people engage. Without knowing how documents. FIELDWORK AND VIDEO DATA Although the primary data for analysis are video recordings of naturally occurring activities. It is also important to become fami- liar with the tools and technologies used by participants and the ways in which the various systems operate and are used within the setting. Each action displays an understanding of the prior. are organi- zed. it is critical that the researcher undertakes more conven- tional fieldwork.

Therefore we attempt to take precautions both to reduce ‘reactivity’ and to assess data for influence of the recording. the placement of the camera can help in this regard too. Goodwin 1981. We routinely leave the camera running and therefore choose a wide enough angle to accommodate basic shifts in orientation and movement by the participants.1 CAMERA ANGLE FOR THE MEDICAL CONSULTATION captured. we positioned the camera in order to capture as much of the face and bodies of both participants as possible. Harper 1994.1 The main reasons for leaving the camera stationary in the setting are to allow the fieldworker to leave the consultation and to ensure that the participants are distracted as little as possible by the recording equipment. Of course. Prosser 1998) are sensitive to our part within and influence on the scene. Most notably. we attempted to select an angle that enabled us to see clearly the objects on the desk in between the doctor and patient (Figure 4. the patients are asked beforehand to participate in the study and thus they are aware that the camera is . Both in undertaking field observation and video recording. we like other field researchers (see e.108 PUTTING THE PRACTICE INTO THEORY FIGURE 4. Goffman’s powerful discussion of participation points to the inevitable significance of an individual within range of an event and in parti- cular the person’s contribution to the interaction (Goffman 1981). For ethical reasons. Additionally. Grimshaw 1982. our placement of the recording equipment was directed by spe- cific concerns. The fairly standard positioning of the doctor and patient around the desk facilitated a relatively constant position. given our interest in bodily as well as spoken conduct. In the case of the medical consultation that we will discuss in detail later.1).g.


filming. However, the field researcher is able to assess where best to locate
the camera so that participants are able to disregard the recording.
After an initial recording phase we often return to the setting for further
fieldwork. For example, over the past few years we have undertaken a num-
ber of projects concerned with the control rooms and operation centres of
London Underground. Following preliminary analysis, we returned to the
control rooms to undertake more focused fieldwork and collect further video
recordings (sometimes choosing different angles to provide access to differ-
ent viewpoints on activities). This iterative characteristic of field studies is
well known, and provides a critical resource not only for developing one’s
understanding of the setting, but also in refining analytic observations and
insights. In certain cases, where we were puzzled by particular events we
returned to the field with a small ‘video Walkman’ to play and discuss
extracts from the video with the participants themselves; their observations
not so much providing analytic resources but helping clarify understanding
of particular incidents, specialized language or technologies.


Transcribing the data

Analysis of the video recordings involves the detailed scrutiny of particular
fragments and we will consider one such fragment, a brief extract from the
beginning of a medical consultation in this section. However, we will first
consider how we initially approach the analysis of any fragment of data. One
of the critical ways in which we can become familiar with a fragment and
begin to explicate the arrangement and organization of the participants’
actions is through the transcription of aspects of the interaction. To do this
we draw on conventional orthographies used for the transcription of talk
which provide a vehicle to begin to come to grips with the details of the talk
and the ways in which it emerges. It allows us to clarify what’s said, by
whom and in what way, and to begin to explore potential relations between
aspects of the interaction. The orthography used within conversation analy-
sis was primarily devised by Gail Jefferson and we have summarized some
of the symbols in the Appendix (see also Jefferson 1984). Transcription does
not replace the video recording as data, but rather provides a resource
through which the researcher can begin to become more familiar with details
of the participants’ conduct.
It is perhaps worthwhile introducing our example at this point. The frag-
ment is drawn from a project concerned with medical practice in primary
health care. It involved extensive fieldwork, discussions with practitioners
and extensive video recording of actual consultations. The fragment
involves the first few moments of a consultation. It gives a sense of how the
consultation begins and the relevance of the various material sources to the
interaction between patient and doctor.


Fragment 1 Transcript 1
((P. enters the surgery))
Dr: Do sit down::
Dr: What’s up?
P: I’ve had a bad eye::: (.) °in there=
Dr: =Oh: yeah

Talk is laid out turn by turn, the length of silences and pauses measured in
tenths of a second and captured in brackets, for example ‘(4.8)’. The colons,
as in ‘down::’, indicate that the prior sound is stretched, the number of colons
indicating the length of the sound. The underlinings, as in ‘up’, indicate that
the word, or part of the word, is emphasized. ‘°’ indicates that the following
word is said quietly, and ‘=’ that the following utterance is latched to the
prior. ‘(.)’ indicates a mini-pause, a pause or silence of two-tenths of a second
or less. Double brackets, ‘((P. enters the surgery))’, house transcribers’
descriptions of actions or events.
There is no general orthography used for the transcription of visual and
tactile conduct, but over the years researchers have developed ad-hoc solu-
tions to locating and characterizing action (see e.g. Goodwin 1981; Heath
1986; Kendon 1990). In our own studies we map fragments developing a
characterization of at least the onset and completion of particular actions
and their relations to each other. This often involves the use of graph paper,
laying talk and silence horizontally across the page, and then mapping the
details of the conduct in relation to each other. We also include notes on the
use of various artefacts and significant changes in information displayed or
documented within the environment (for example in records, on screens and
display boards and the like). Consider the example shown in Figure 4.2. It is
a version of one of the original ‘maps’ developed for the fragment under dis-
cussion. The transcript presents the participants’ conduct horizontally, with
dashes capturing the length of silences and pauses, one-tenth of a second
indicated by one dash.
These more detailed transcriptions of a fragment are simply devices to
enable the researcher to identify particular actions and to preserve a rough
record of what has been found at some particular stage of the analysis. They
are not designed to be read or used by others, or of course to provide a
literal or true characterization of the events. However, they do provide a
critical resource to help the researcher to establish the range and complexity
of conduct within a particular fragment, and with which to begin to identify
its character and location. Indeed, without logging the details of a fragment
in this or a similar fashion, it is found that conduct is frequently mislocated,
mischaracterized and in some cases missed all together. Transcription
provides a vehicle for clarifying the location of actions and in exploring the
potential relations between co-occurring and surrounding talk and bodily



conduct. In particular, it is only through detailed investigation of the location
and character of particular actions that we can begin to clarify their emergent
and contingent relations between participants’ conduct; the sequential char-
acter of action which is the pervasive organizational feature of human activ-
ity in social interaction.
The situated character of practical action, and the interest in the methodo-
logical resources used by the participants themselves, inevitably drives ana-
lytic attention towards the investigation of activities and events within the
contexts in which they occur. Detailed and repeated inspection of the accom-
plishment of actual activities, coupled with the analytic orientations briefly
discussed above, provide resources through which researchers can begin to
identify the practices and reasoning through which particular events are
produced and rendered intelligible.
In considering the fragment, for example, we would then want to develop
a characterization of the activity which has evolved through our close look-
ing in generating the transcript. To illustrate consider our description of this

Developing an analysis

As the patient enters the consulting room and walks towards the chair
alongside the desk, the doctor utters ‘What’s up?’ The utterance invites the
patient to deliver his reason for seeking professional help. It projects a
sequentially relevant action for the patient, and following a few seconds’
silence, the patient does indeed deliver the appropriate response, ‘I’ve had a
bad eye::: (.)°in there.’ The exchange involves the transition of the consulta-
tion from the ‘preliminaries’ to the ‘business at hand’, and the patient’s
response provides resources for subsequent enquiries, diagnosis and treat-
ment. Progression into the business of the consultation is also dependent
upon the bodily conduct of the participants and the use of particular tools
and artefacts. It is worth noting at the outset for example that the patient’s
reply is delayed by four seconds or so, and is then accompanied by a gesture.
A second transcript, capturing particular aspects of their visual conduct,
may be helpful. In this case the detailed ‘map’ depicted earlier has been pared
down to leave those aspects most critical to the analysis. The transcript is
accompanied by descriptions of particular actions or events (Figure 4.3).
The doctor produces the initiating utterance as the patient crosses the
room. The patient sits down within a second or so, but remains silent. The
doctor reads the medical records as the patient sits down. A few moments
later, the doctor turns from the records to the patient. The patient immediately
turns to the doctor and begins to reply. As he replies, he gestures towards his
eye, and the doctor moves forward and inspects the difficulty.
The patient’s reply is sensitive to the doctor’s use of the medical record.
He withholds his reply, the sequentially relevant response, until the doctor
turns from the record to the patient and thereby visually completes the activity

P. crosses sits

room in chair

What’s up? – – – – – – – – –, – – – – – – – – –, – – – – – – – – –, – – – – – – – – –, – – – – – – – – – I’ve had a bad eye:::

Dr turns from

reads records the records to P.




in which he is engaged. The doctor’s lack of orientation to the patient is
legitimized by virtue of his reading the patient’s medical record, as informa-
tion gleaned from the record may be relevant to the subsequent proceedings.
It should be added that glancing through the last few entries of a patient’s
record is a recurrent feature of the beginning of medical consultations at least
in general practice. It provides information to enable the doctor to know, for
example, whether the patient is returning to discuss a particular complaint,
or whether the patient has suffered serious difficulties in the past which
might be relevant to the presenting complaint.
There is further evidence to suggest that the patient’s conduct is sensitive
to the doctor’s use of the medical record, and indeed, may encourage the
doctor to bring the activity to a quick conclusion.
The patient glances at the doctor as he sits down, a juncture within the
developing course of his own activity at which it may be relevant to reply
and set the proceedings in motion. The doctor is reading the record. A
moment later, the doctor raises his right hand and smoothes the page of the
record. The gesture differentiates the doctor’s activity. It displays a shift in
alignment towards the document and potentially projects the possible com-
pletion of the reading activity by arranging the document as if in readiness
for writing. As the hand presses the page, the patient turns to the doctor. He
opens his mouth as if beginning to speak. As he opens his mouth, the doctor
turns to another area of the page, while his hand moves forward to clasp a
date stamp. The patient closes his mouth, turns away and licks his lips
(Figure 4.4).
The patient therefore is sensitive to the ways in which the doctor reads
and manipulates the medical records. The very lack of orientation by the
doctor is accountable by virtue of his use of the record, and reading the
record can be seen as relevant to the consultation and movement from
the ‘preliminaries’ to the business at hand. Similarly, for the doctor, the
patient’s lack of immediate response, and his shifting orientation in the pro-
ceeding silence, is sensible by virtue of the patient’s sensitivity to the use of
the records. The medical records therefore are an integral feature of the parti-
cipants’ activities, both in the ways in which they produce their conduct, and
in how they make sense of each other’s actions.
One can also gain a feel for the emergent and contingent character of the
participants’ conduct. The patient’s actions are sensitive to the emerging use
of the record, they differentiate the actions of the doctor moment by moment
within the developing course of the reading. Where the doctor looks, how he
scans the page, his raising of the hand, the pressing of the page, inform the
ongoing production of the patient’s actions, just as the doctor himself is sen-
sitive to the patient’s shifting orientation to the activity. We can see therefore
how the ‘situated’ and contingent character of practical action is shaped and
created through the moment-by-moment production of the participants’
actions. Each action is sensitive to the actions of the other, and provides the
basis for subsequent conduct, as they emerge within the developing course
of the activity.

P. sits turns to Dr & looks turns to

in chair takes in-breath down Dr

What’s up? – – – – – – – – –,– – – – – – – – –,– – – – – – – – –,– – – – – – – – –,– – – – – – – – –,– – – – I’ve had a bad eye:::

Dr flattens page reads turns from

reads records & reorientates page record to patient




In this way therefore we can begin to disassemble aspects of the social and
interactional organization which feature in the accomplishment of a particu-
lar event and provide for its character and uniqueness. Even this cursory
glance at the fragment begins to reveal the complexity of the participants’
activities and the resources which are brought to bear in the production and
intelligibility of the beginning of the consultation. It reveals the emergent
character of the participants’ actions and the ways in which they are inter-
actionally organized and accomplished. In the case at hand, it can be seen
that while visual and tactile aspects of the participants’ conduct are not
organized on the turn-by-turn basis characteristic of talk, nonetheless the
sequential character of conduct is a critical property of the production of
action and its intelligibility. For example, the elicitation of the doctor’s gaze
is sequentially responsive to an action by the patient, and the doctor’s reori-
entation forms the foundation to the beginning of the consultation. In turn-
ing to consider the methodological resources used by the participants in the
activity’s production and intelligibility therefore, attention inevitably turns
to consider the ways in which actions are interactionally coordinated
moment by moment, step by step. In this sense therefore context and intelli-
gibility of the action is unavoidably and continuously emergent and assem-
bled from within its production.


This analysis is designed to reveal how close looking at the details of inter-
action can be used to uncover critical resources used by participants in the
organization of the medical encounter. As Hughes (1958) and many other
ethnographers suggest, social interaction forms the foundation to the pro-
duction of everyday activities, whether they involve a medical consultation,
the operation of a control centre or a visit to museum or gallery (see e.g.
Heath and Luff 2000; vom Lehn et al. 2001). Despite the recognition of the
importance of interaction to everyday life and social organization within
qualitative sociology, it has remained largely disregarded, ‘noticed but not
seen’, remarked upon but to some extent unexplicated. It is not surprising
however that ethnography, at least in sociology, has largely been unim-
pressed by the growing body of research concerned with language use and
in particular talk in interaction. Such studies while providing impressive
insight into the socially organized character of talk have seemed to disregard
an array of considerations and concerns that form the focus of ethnographic
inquiry. In this chapter, we wish to suggest that video-based field studies
coupled with an appropriate analytic orientation, namely ethnomethodo-
logy and conversation analysis, can provide a vehicle with which to address
one or two of the issues and substantive concerns that form part of more
traditional ethnographic research, even though the phenomena of interest,
the conceptual orientation and the ability to legitimize ‘broader’ contextual
characteristics may seem somewhat over constrained.

artefacts and the like. but also through the ways in which they are used. this single turn at talk. talk-in- interaction. The doctor’s ‘lack’ of orientation towards the patient is not simply ‘withholding gaze’. We too are keen to include the material environment. is often treated as forming the frame- work for action and in various ways providing the resources including symbolic representations for organization and interpretation of action (see e. but also designed with regard to the accompanying bodily action. Consider for example the patient’s reply. The utterance is not only occasioned by the doctor’s visual and vocal conduct. the patient’s own conduct is sensitive to the ways in which these artefacts are used. and even how the particular use of an artefact can prospectively display what it will take for it to be complete. but rather is sensible by virtue of his reading of the patient record and his use of the date stamp. As we have seen. within the analytic scheme. we can begin to examine the ways in which talk is inextricably embedded within the participants’ visual and tactile con- duct. and of course bodily as well as spoken conduct. the chair and its particular use marking the completion of a relevant course of action. objects. among other characteristics. noticed. The patient’s reply and gesture render the body relevant. just as the visibility of a blemish in the eye provides sense to the utterance. Blumer 1962). then and there. The response emerges in regard to the doctor’s bodily conduct and in particular his visual realignment which enables him to see and look at the patient. and reflexively constitutes. and constitutes the relevance of the body and its conduct. seen. Not only is the participants’ conduct oriented to these objects and artefacts in particular ways at particu- lar moments. their bodily conduct and the local ecology of objects and artefacts. For example. These objects and artefacts come into play within the developing course of action. and its significance then and there within the interaction. Rather however than treating material realities as having an overarching influence on the field of conduct and thereby assuming that their sense and significance remains stable throughout the emerging course of events. and in particular the ability of the doctor to see as well as hear about the source of the complaint. In this. disregarded and the like. the particular object gains a specific sense and relevance from within the course of action. and a myriad of other ways.g. ANALYSING INTERACTION 117 Taking the example at hand. the initial delay in the patient’s response is produced and accountable with regard to the patient sitting. in a particular way. his reason for seeking professional help. In undertaking video-based field studies of social interaction we have the opportunity of addressing characteristics of action and of settings which have formed a concern for more traditional ethnography. gains its sense and significance by virtue of the ways in which it invokes an aspect of the body. The spoken and bodily conduct of the participants is inseparable from. material features of the local environment. and invite the doctor both to hear about and to inspect the complaint. The utterance. and thereby serve to mark a potential turning point in the beginning of the consultation. such . The physical envi- ronment. is accomplished through.

seen. Heath and Hindmarsh 2000. and they gain their sense or meaning at those moments from within the action in which they are momentarily ren- dered relevant. Such studies are not only beginning to provide a distinctive contribution to our understanding of organizational activities. and ethnomethodology and con- versation analysis provide resources through which we can begin to unpack the interactional organization of activities and events in these complex techno- logical settings and demonstrate the relevance of environment to actual courses of action. If you like. news rooms. for particular purposes. our problem is not simply taking the material environment seriously (like other potentially rele- vant features such as the organizational setting. They feature both in the production of action and the ways in which the participants make sense of each other’s conduct. on monitors. provide unprecedented access to such complex tasks. . noticed. Video recordings. used. interaction and technology and provides programmatic examples of the ways in which video-based field studies of interaction can bear upon topic and concerns ordinarily associated with more traditional ethnography. This corpus of research. feature in action and social interaction. 2000). Goodwin 1995. This concern with the ways in which the material environment features in practical action and interaction is reflected in the growing body of empirical research concerned with tools and technologies in complex organizational environments. often using multiple cameras. there is a vast array of information provided in documents. referred to.118 PUTTING THE PRACTICE INTO THEORY as a medical consultation. and in the ways in which participants themselves recognize and make sense of each other’s conduct. Hindmarsh and Heath 2000). These stud- ies consist of analyses of work and interaction in settings such as control cen- tres. and the like. we need to examine the ways in which objects. in and through the partici- pants’ action and interaction (see also Wootton 1994. but also. augmented by extensive fieldwork. As we have seen. the critical issue is exploring and demonstrating what is relevant and how it is constituted as relevant. material features of the immediate setting are invoked. banks. The immedi- ate ecology of objects and artefacts provides resources for the production of action. artefacts and the like come to gain their particular significance at specific moments within courses of action. In settings such as control rooms. The sense and determination of the material environments is reflexively constituted. across diagrams at any one time. commonly known as ‘workplace studies’ (see for example Luff et al. at particular moments. consists of naturalistic studies of work. such settings provide a substantive demonstration of the issues raised throughout this chapter. but also chang- ing the ways in which the social and cognitive scientists conceive of the inter- action between human beings and technologies such as computers. Not only do they powerfully illustrate the importance of taking material features of the environment such as tools and artefacts seriously. within action and interaction (see for exam- ple Heath and Luff 2000). Streeck 1996. participants’ background and the like) but rather analytically demonstrating how such characteristics become relevant and reflexively constituted in action. in addressing the ways in which such tools and the complex array of information they provide.

in which the patient is the speaker: 30 P: in fact I’ll do it right now. It is normally necessary to use separate microphones and we mainly use a multi-directional microphone PZM. We always make and work on copies of the original data. (0. not necessarily a question ∧ Marked rising shift in intonation ∨ Marked falling shift in intonation °hhh In-breath hhh Out-breath Overlapping utterances are marked by parallel square brackets: e. APPENDIX: CONVENTIONS FOR TRANSCRIBING TALK The Identity of the speaker is indicated in the margin. Continuing intonation.g. 30 P: in fa [ct I’ll do it right now.) A pause which is noticeable but too short to measure = No discernible interval between adjacent utterances erm::: Elongated sounds – the longer the elongation. the more colons are added to the utterance or section of the utterance but. sometimes alongside a line number. We primarily use Sony EVC 500 machines for analysis. they have stable still frame and reasonable slow motion. ANALYSING INTERACTION 119 NOTE 1 We are currently using both analogue and digital video equipment. A stopping fall in tone. okay .An abrupt cut-off of an utterance or part of an utterance is marked by a dash under Emphasized stretches of talk are underlined >right< Faster stretches of talk °that° Quieter stretches of talk . not necessarily the end of a sentence . 31 D: [oh right.6) A pause timed in tenths of a second (. but in certain settings with significant background noise a mono-directional microphone such as a Sennheiser MKE300 is more successful. and digital. not necessarily between clauses of sentences ? Rising inflection. This example shows line 30 of a transcript. The analogue cameras are Sony Hi8 TRV65E. as the repeated replaying necessary during transcription and analysis can severely damage tapes. Sony TRV900E.

F. Chicago: Aldine. H. Heath. E. C. (1963) Passage through Crisis: Polio Victims and their Families. E. (1992) ‘Veiled morality: notes on discretion in psychiatry’. Rose (ed. in A. E. (1984) Garfinkel and Ethnomethodology. Goffman. New York: Free Press. London: Routledge and Kegan Paul. D. New York: Doubleday. J. Cambridge: Cambridge University Press. (1967) Interaction Ritual. Goodwin. (2000) Technology in Action. J. (1981) Conversational Organisation: Interaction between a Speaker and Hearer. J. Goffman. C. (1995) ‘Seeing in depth’. Goffman.D. Grimshaw.120 PUTTING THE PRACTICE INTO THEORY REFERENCES Atkinson. (1995) Medical Talk and Medical Work: The Liturgy of the Clinic. (eds) (1991) Talk and Social Structure. Heritage. and Luff. A. Indianapolis. Garfinkel. Becker. (1983) ‘The interaction order: American Sociological Association. (1961) Asylums. Drew and J. P. Heritage.H. London: Academic Press. Davis. London: Sage. E. Cambridge: Cambridge University Press. D. and Hindmarsh. Sociological Methods and Research. Thousand Oaks. .K. Social Studies of Science 25(2): 237–274. in N. D. Cambridge: Cambridge University Press. Bergmann. P. Englewood Cliffs. Culture and Activity 7(1/2): 81–104. B. NJ: Prentice-Hall. in D. (1994) ‘On the authority of the image: visual methods at the cross- roads’. (1981) Forms of Talk. Cambridge: Cambridge University Press. (2000) ‘Configuring action in objects: from mutual spaces to media spaces’. (1982) ‘Sound-image data records for research on social inter- action’. in P. P. Lincoln (eds) Handbook of Qualitative Research.H. and Heritage. Boden. (1984) ‘Introduction’. Goffman. and Zimmerman. H. New York: Free Press. Oxford: Basil Blackwell. (1965) Awareness of Dying. C. J.R. in P. Zimmerman (eds) Talk and Social Structure: Studies in Ethnomethodology and Conversation Analysis. ten (1991) ‘Talk and institution: a reconsideration of the “asymmetry” of doctor–patient interaction’. Goffman. CA: Sage. Oxford: Polity. Denzin and Y.C. P. and Atkinson. C. IN: Bobbs-Merrill. (1963) The Outsiders. (1967) Studies in Ethnomethodology. and Strauss. Atkinson and J. Heath. Heath. (1962) ‘Society as symbolic interaction’. New York: Anchor. H. Goodwin. (1986) Body Movement and Speech in Medical Interaction. Mind. Cambridge: Polity. E. Have. Paris). 1982 Presidential Address’. Drew. Blumer. 179–192. Heritage (eds) Structures of Social Action: Studies in Conversation Analysis. Heritage (eds) Talk at Work: Interaction in Institutional Settings. J. (1963) Behaviour in Public Places: Notes on the Social Organization of Gatherings. pp.S. Glaser. A. Harper. American Sociological Review 48(1): 1–17. (eds) (1992) Talk at Work. Boden and D. Cambridge: Polity. Special Issue on Sound-Image Records in Social Interaction Research 11(2): 121–144. P. Cambridge: Cambridge University Press (Editions de la Maison des Science de l’Homme.) Human Behaviour and Social Processes. C.

Jefferson. Drew and J. IN: Bobbs-Merrill. Schegloff. (1992) ‘On clinicians co-implicating recipients’ perspective in the delivery of diagnostic news’. Roth. J. Sarangi and C.) (1998) Image-based Research: A Sourcebook for Qualitative Researchers. Heritage (eds) Talk at Work: Interaction in Institutional Settings. IL: Free Press. Journal of Child Language 21: 543–564. Oxford: Blackwell. Sacks. knowledge and competence in the pharmacist/client encounter’. Pilnick. in J. (2001) ‘Exhibiting interaction: conduct and collaboration in museums and galleries’. (1994) ‘Object transfer.J. Mediation. (1971) The Sociological Eye. E. Language 50(4): 696–735. Purdue University. (1974) ‘Opening up closings’. R. C.. (1950) The Sociology of Georg Simmel. intersubjectivity and third position repair: early developmental observations of one child’. A. (1974) ‘A simplest systematics for the organization of turn-taking for conversation’. ANALYSING INTERACTION 121 Hill. London: Routledge and Kegan Paul. (1971) ‘The everyday world as phenomenon’. E. G. Glencoe.A. D. in P.) Ethnomethodology: Selected Readings.J. (1990) Conducting Interaction: Patterns of Behaviour in Focussed Encounters. J. (1996) ‘How to do things with things: objets trouvés and symbolization’. (2000) Workplace Studies: Recovering Work Practice and Informing System Design. G.D. Work and Institutional Order: Discourse in Medical. A. Atkinson and J. Chicago: Aldine. vom Lehn. D. London: Falmer. Streeck. Human Studies 19: 365–384. Journal of Contemporary Ethnography 29(5): 517–556. edited by K.H. Heritage (eds) Structures of Social Action: Studies in Conversation Analysis. H. (1964) Psychiatric Institutions and Ideologies. (1958) Men and their Work. and Sacks. IL: Free Press. H. Hughes.C. Glencoe. Glencoe. Harmondsworth: Penguin. Social Psychology Quarterly 61: 301–320.S.W. D. E. (1963) Timetables: Structuring the Passage of Time in Hospital Treatment and Other Careers.C. Institute Monograph Series no. Luff. Cambridge: Cambridge University Press. J. P. Prosser. Cambridge: Cambridge University Press. London: Routledge and Kegan Paul. Zimmerman. M. (1999) ‘Warriors or collaborators: reworking methodological con- troversies in the study of institutional interaction’. Cambridge: Cambridge University Press.. in J. Maynard. Hindmarsh. Silverman. Indianapolis. Peräkylä. Turner (ed. Symbolic Interaction 24(2): 189–216. (1998) ‘Authority and intersubjectivity: the delivery of diagnosis in primary health care’. 1. Heath. Berlin: Mouton de Gruyter. C. (2000) ‘Sharing the tools of the trade: the interactional constitution of workplace objects’. Strong. and Pollner. and Heath. (1984) ‘Transcript notation’. D. E. P. (ed. J. J. C. in R. A. and Crittenden.) Understanding Everyday Life.M. K.L. A. A. and Heath. and Hindmarsh. (eds) (1968) Proceedings of the Purdue Symposium on Ethnomethodology.A. Wolff. Wootton. Sociology of Health and Illness 20(1): 29–51. Strauss. in S. and Management Settings.. J. and Jefferson. Simmel. Volumes I and II. Roberts (eds) Talk. IL: Free Press. H. Schegloff. Institute for the Study of Social Change. G. Douglas (ed. (1998) ‘“Why didn’t you just say that?” Dealing with issues of asym- metry. (1979) The Ceremonial Order of the Clinic. Hindmarsh. Cambridge: Cambridge University Press.M. Sacks. (1992) Lectures in Conversation. Kendon. Hughes. .




Analogously for the artist the appreciation and understanding of (say) Sisley’s painting of the riverbank at St Mamnés cannot be extended to statements about the specifications of riverbanks elsewhere.1 My argument will be that generalizations are legitimate if they are treated and made explicitly as (what I term) moderatum generalizations. where the aim is to produce the social equivalent of a Sisley painting. for sociologists to interpret a set of bio- graphical outcomes in a single agent. Policy-makers need data they can rely on to reach their decisions. I will further argue that an affirmation of the possibil- ity of generalization is an acknowledgement of the limits of interpretivism. it is said. An ideographic account is an interpretation by a situated agent (the researcher) of a never to be repeated event or setting. as part of an explanation–prediction schema in science. It follows from this that the particulars of such accounts of the social world therefore cannot be used to generalize to other instances. or to evaluate existing programmes. Likewise. In sociology the question of generalization in interpretive research lies at the heart of an old debate about whether the discipline is nomothetic or ideo- graphic in character – broadly speaking whether it is ‘scientific’ or ‘human- istic’. in the form of focus groups or in-depth interviews. As Ray Rist (1998) put it: . is called upon to provide just such evidence in the creation of local or national policy. This produces a problem for interpretive research. within the context of a pluralistic approach to research. autonomous nature of human subjects produce states which are not amenable to the explanation–prediction schema of natural science. the self-reflective. then of course generalization.5 GENERALIZATION IN INTERPRETIVE RESEARCH Malcolm Williams In this chapter I will consider the question of whether it is possible to generalize from the results of interpretive research. or the social milieu of group of agents. Increasingly interpretive research. it is intended as a unique appreciation of a moment. is permissible. but it won’t do in circumstances where decisions affecting the emotional and material existence of people are made on the basis of evidence produced. Such a relativity of perspective is undoubtedly fruitful in a humanistic contemplation of the social world. at a given time is a unique act. Yet many of those working in the interpretivist tradition have often cast doubt upon whether this is the case and indeed the method itself is adopted because. An ideographic disci- pline concerns itself with the understanding of an instance in a unique context. If it is the former.

Having outlined a pos- sible form that generalization might take in interpretivism. or individual biographies. This denial. maintained or destroyed and must therefore look to evidence manifested in the specific fea- tures of the social world to do so. we quite often need to know the ethnographic basis of the statistics pro- duced by the survey (Gephart 1988). Moreover survey research cannot usu- ally provide the contextual detail necessary to interpret even its own results. I will suggest. Moreover it is not just an issue for the prag- matic investigation of social problems or policy choices. Sensitive topics or difficult to research popu- lations (for example) can be known only through interpretive methods (see for example Renzetti and Lee 1993). INTERPRETATION AND ANALYSIS What are the contours of the issue? Is the problem or condition one that is larger now than before. but by no means all. some of the foregoing could be obtained through survey research. areas. or institutions experience this condition now as earlier? …. If interpretivism is to be any use at all in social policy formulation or eval- uation it must be able to say something authoritative about instances beyond the specific ones of the research.126 GENERALIZATION. or smaller? Is anything known about whether the nature of the condition has changed? Do the same target populations. Finally I will con- clude that a recognition of such limits and possibilities entails the prospect of a reconciliation between the ideographic and nomothetic traditions. about the same. the sociologist is interested in knowing something of the social world beyond isolated and fragmented accounts of interactions. I will provide some illustra- tions of the manifestation of generalization in interpretive research and the denial of its possibility by methodologists. is based on a narrow definition of what it means to generalize. is a topic that is at least visually familiar to many of us in con- temporary western societies. Such features are quite often at the level of micro-detail. The first study. or do not. but must be understood in such a way as to explain how they do. reproduce or destroy social structures. If one looks to the history of sociology. in the second section of the chapter I discuss its limits and possibilities. create. In this chapter I will proceed as follows: first. What are the different understandings about the condition. of children and fruit machine gambling (Fisher 1993). it is precisely at the level of micro-detail that researchers have claimed to find evidence of structure and indeed many of the classic interpretivist studies in sociology are attempts to explain particu- lar forms of social structure. DOING IT AND DENYING IT Doing it I offer two brief examples to illustrate the kinds of claims that are made in interpretive research. It can be seen as typical of the many projects . The sociolo- gist wants to know about how social ‘structures’ are created. its causes and effects? (Rist 1998: 406) Now of course.

The ‘typicality’ of the location and the subsequent findings are simply assumed. but the evidence itself has a dif- ferential status which is not discussed. Rent-a-Spacers. Fisher’s observa- tions led her to construct a typology of five categories of fruit machine player: Arcade Kings. an examination of juvenile fruit machine gambling demonstrates ‘the diversity of primary orientations which are intrinsic to gambling’ (1993: 470). or the way in which indi- viduals used the social space of the arcade. Fisher’s investigative strategy. maintains that ‘musical and stylistic sensibilities exhibited by the young people involved in the dance music scene are clear examples of a form of late modern “sociality” rather than a fixed subcultural group’ (Bennett 1999: 599). Action Seekers and Escape Artists (1993: 458). Thornton (1995) herself claims to have conducted observations at over 200 discos. Nevertheless she doesn’t tell us why the particular arcade was chosen. while Escape Artists gamble primarily to escape personal problems. if not the only one. Machine Beaters. GENERALIZATION IN INTERPRETIVE RESEARCH 127 aimed at testing or developing theoretical constructs that have potentially important sociological or policy implications. whereby inductive support for one’s argument is drawn variously from theoretical positions (in this case Maffesoli 1996). nor the basis upon which respondents were selected. Indeed Sue Fisher’s work. The typologies describe strategies of playing. The author. average. why she conducted the number of interviews she did. Andy Bennett (1999). Machine Beaters were individualists who liked to play alone with the aim of ‘beating the machine’ (1993: 462–463). majority or main- stream’ (1995: 106). For example. other interpretive studies (here Thornton 1995). second-order summaries of other studies (here Shields 1992).2 as well as original research. has enormous importance in helping us to understand an actual or potential social problem. Rent-a-Spacers primarily used the arcade as a meeting place. was probably the best available in that micro-level detail was required to establish such heterogeneity. but this time is concerned with youth style and musical taste. In a society where juvenile gambling is considered to be a social problem a study such as this allows us to see that gambling is not an homogeneous concept. Her study was based upon an unobtrusive participant-observation of children in a gaming arcade (she worked as a part-time cashier in the arcade). Conversely Bennett claims that ‘for many enthusiasts . Additionally she conducted ten in-depth interviews with young fruit machine players and four group inter- views with secondary school children (Fisher 1993: 454). The second study is also about the social milieu of young people. raves and concerts between 1988 and 1992 and maintains that ‘I was unable to find a crowd I could comfortably identify as typical. The position Bennett advances is quite a plausible one. The implication is that strategies to discourage such practices must arise from a consideration of its hetero- geneity. ordinary. clubs. the meanings that players attached to their activity. The article itself uses a style of argument currently popu- lar in interpretive sociology. Fisher concludes that while existing sociological explanations of gambling are heuristically useful. and a great many studies of this kind.

Though the distinction and ensuing debate between the nomothetic and ideographic approach to studying the social world has its origins in the methodenstreit dispute of the nineteenth century. in this respect there is no logical difference between the work of an interpretive researcher conducting detailed observations of a social setting and a large- scale national survey. Many of the issues about the status of interpretivism can. led him to both criticize and endorse scientific method. metaphorically speak- ing. few interpretive accounts ever do discuss such things.3 While trying to bridge the gap . Their work. is not an appreciation of the riverbank at St Mamnés. In both they are inferring from specific instances to the characteristics of a wider social milieu. but rather that his work typifies the longstanding angst in interpretivist method- ology about the epistemological status of the method. Both wished to use the evidence to support a sociological argu- ment that the world is a particular way. to capture the nuances and the singular characteristics of the social environment. however speculatively or tentatively held. Though the evidence in both cases supports Bennett’s argument. Of course it may be that Bennett’s conclusion is based upon a great deal more observation than he reveals and it is also possible that his informants are very much more typical (in their untypicality) than many of Thornton’s informers. temporal experiences’ is ‘clearly illustrated’ (1999: 611) on the basis of one conversation with clubbers in Newcastle. at the same time as advocating a ‘naturalistic’ approach to investigation. Denying it Both Fisher and Bennett make generalizing claims. whether it is. as Martyn Hammersley (1989) documents. such debates have contin- ued to preoccupy methodologists for much of the twentieth century. In each study the researcher attempts to interpret what is going on accord- ing to the subjective frame of reference of those observed.128 GENERALIZATION. to seek explanations. The failure of a lot of empirical research to engage with these questions does not necessarily indicate poor quality research. INTERPRETATION AND ANALYSIS “clubbing” appears to be regarded less as a singular definable activity and more as a series of fragmented. the status of each set of evidence is at least quantitatively different. Nevertheless to investigate is to ask ‘why’ questions. but is a stronger claim about the way riverbanks are! Of course their explanations may or may not be ‘right’ or ‘correct’ and indeed both researchers would probably stress the fallibility and incompleteness of the generalizations they make. but rather might itself be seen as an outcome of unfinished methodological business in interpretivism – specifically. Indeed this is not unusual. This is not to suggest that Blumer is somehow the author of the present situation. yet neither acknowledge nor deny they are doing so. Blumer’s contradic- tory statements about the possibility of a social science. be found in the work of Herbert Blumer. can or should be a nomothetic discipline.

nevertheless aimed to uncover processes in social life (Hammersley 1989: 162). On the one hand there have been attempts to develop the kinds of non-statistical. Alan Bryman’s (1988.) with concern more for the authenticity of contextual understanding. reliability of instruments etc. The post-Kuhnian anti-positivist revolution of the 1960s and 1970s was not conducive to a fruitful discussion of issues such as generalization in interpretivism. Blumer. is a fractured one. While survey researchers could agree on a number of key tenets to the approach (e. Denzin (1983: 133). though even here this was a resource for most approaches. particularly that of the status of interpretive knowledge. labelling theory. It remains. conflict the- ory. the first from Norman Denzin (1983) and the second from Guba and Lincoln (1982). maintains that every instance of human interaction represents a . Thus the possibility of generalization is rarely denied outright and where it is denied. 1998) contention that there is a disjunction between epistemology/methodology and method seems to be correct. of course. have not been carried forward in a great deal of empirical research and as Hammersley (1989: 220) concludes. GENERALIZATION IN INTERPRETIVE RESEARCH 129 between nomothetic and ideographic approaches. the centrality of probability. ‘there has been a tendency in recent years to ignore the problems. Instead methodological concerns have usually since translated into a search for ‘interpretive valid- ity’ of one sort or another (validity as culture. which while emphasizing the indeterminism of human action. To discuss gener- alization at all was to play by the positivists’ rules. that the adoption of interpretivist methods has quite often been the result more of adherence to an anti-science position than the utility of the method itself. The methodological priorities and messages were very different in each and often incompatible. as language etc. that with a few honourable exceptions the methodo- logical issues raised by Blumer’s work. he alternately leaned toward each. critical theory etc.g. however. in claiming that ‘the interpretivist rejects generaliza- tion as a goal and never aims to draw randomly selected samples of human experience’. but became a topic itself for ethnomethodology (Williams and May 1996: 73–76). One might argue. other than the importance of meaning. yet rigorous scientific approach to the method Blumer favoured. as ideology. as I have elsewhere (Williams 2000: 87–103). Thus a requirement must be that the sociologist move from interpretations of spe- cific agents’ actions and situations. to social explanation. was in the symbolic interactionist tradition. In order to throw the issue into the sharpest light it is worth looking at two of the outright denials. and to forget the diversity of views adopted by both advocates and critics of qualitative method’. On the other hand there has been a plethora of cross-influences with ethnomethodology. as gender.) there was little the various strands of interpretivism could agree on. caveats are often entered. because as Altheide and Johnson (1998) point out the ‘traditional criteria of methodological adequacy were formu- lated and essentially “owned” by positivism’ (1998: 286). in interpretivism. Indeed in the case of a great deal of interpretivism. in analytic induction (which I will discuss below) and grounded theory. Blumer’s legacy.

INTERPRETATION AND ANALYSIS ‘slice from the lifeworld’ (Denzin 1983: 134). This knowl- edge is best encapsulated in a series of ‘working hypotheses’ that describe the indi- vidual case. If the conditions specified by Denzin are what are required for generalization. where for example the concept of ‘fitness’ or ‘adaptability’ can change over short spaces of time and between places (Kincaid 1996: 60). Thus S1 is not a copy of S but an instance of a general deterministic law that governs S also. Denzin attributes.nor context-free’. (Guba and Lincoln 1982: 238) If the foregoing was right then generalization in interpretive research would indeed be impossible.5 Such generalizations are in fact axioms and do no more than express instances of particular laws. the following position: references to the social world that could not be verified under quantifiable. scientifically controlled conditions must – following Wittgenstein’s dictum – ‘be passed over in silence’.130 GENERALIZATION.4 This would rule out generalization in complex biological systems.nor context-free. observ- able. (Denzin 1983: 132) Such a position is probably one unrecognizable to even the most militant logical positivists and the generalizations that would emerge from the vari- ant of ‘positivist’ sociology that Denzin paints could be made only in the ‘closed’ system of the laboratory. In rather similar vein Guba and Lincoln (1982) maintain that The aim of inquiry is to develop an ideographic body of knowledge. How can this be? The problem may well lie in what is meant by generalization. to those he calls ‘positivists’. Open systems could not be studied and probabilistic conclusions would be invalid. Thus the rate of cooling of an electric element is an instance of (and calculable through) the second law of thermodynamics. Conversely different actions can arise out of similarly expressed meanings. One begins to get the impression that what Denzin (1983) and Guba and Lincoln (1982) mean by ‘generalization’ does not fully express all the avail- able meanings. Therefore there is an ‘inherent indeterminateness in the lifeworld’ (Denzin 1983: 133) which leads to too much variability to allow the possibility of generalization from a specific situation to others. then I would have to agree with him that it is impossible in sociology and probably in quite a lot of the natural sciences too! Guba and Lincoln (1982) similarly assert that ‘Generalizations are impos- sible since phenomena are neither time. Let me suggest three possible meanings (there may be more): Total generalizations: where situation S1 is identical to S in every detail. carrying layered meanings which come in multiples and are often contradictory. Generalizations are impossible since phenomena are neither time. . but we have already seen that interpretivists do gen- eralize. The core of Denzin’s argument can be summarized as individual consciousnesses are free to attach different meanings to the same actions or circumstances.

LIMITS AND POSSIBILITIES OF GENERALIZATION IN INTERPRETIVIST RESEARCH To disagree with the rather stark position of Denzin (1983) and Guba and Lincoln (1982) above is not to embrace one which denies the problems raised . This is simply the rela- tionship between sample and population and is the basis upon which most generalizations (other than some in physics and chemistry) in the natural sciences are made. This is the form of generalization made in interpretive research. while they may assume particular forms of significance for clubbers. consists of complex phenomena. Which of these does Denzin (1983) and Guba and Lincoln (1982) say are impossible? Well. It is. GENERALIZATION IN INTERPRETIVE RESEARCH 131 Statistical generalizations: where the probability of situation S occurring more widely can be estimated from instances of s. chemistry and their cognate disciplines. they are the generalizations of every- day life. it seems clear that the first is impossible in the social world and this seems to be trivially true. either knowingly or unknowingly. is make moderatum generalizations. while possible in the social world. The sociological survey usually depends upon some form of probability sampling. of course. Bennett’s assertion that such fluid and eclectic forms of music consumption. whereby within each stratum every case will have the same probability of selection. Outside of the axiomatic laws of nature specified in physics. The second type of generalization. Moderatum generalizations: where aspects of S can be seen to be instances of a broader recognizable set of features. Much of the world. even the complex phenomena of open systems lend themselves to statistical prediction to vary- ing degrees of success (see for example Casti 1991). for which no laws (or at least only statistical laws) are available. On the first two counts then. cannot usually be made from the kind of data generated by interpretive research. 1975). Finally moderatum gener- alizations in their simplest form are the basis of inductive reasoning in what Schutz (1972 [1932]) called ‘the lifeworld’. the researcher is able to statistically express the level of confidence she has that her sample represents the population. natural and social. are not in fact restricted to urban dance music clubs but are also central to other aspects of youth and youth culture (Bennett 1999: 612) is an example of such a general feature capable of reworking and enriching through specific instances of music consumption. total generalizations are impossible. Importantly however. though sometimes (and this is more the case in the natural sciences) alternative means of inference and counting are used (see Bishop et al. but what interpretivists do and I would suggest they are right to do. the gen- eralization of interpretive findings is impossible. But of course. the same basis upon which survey researchers in sociology generalize. The survey researcher mostly depends upon the possibility of drawing the sample from a sampling frame.

Holdaway 1982).132 GENERALIZATION. elusive or the research must be covert (see for example studies by Fielding 1982. Survey researchers wishing to make inferences about a population will decide which are the appropriate sampling units and (as I noted above). that interpretive methods are being used may be because the population in question is rare. as she points out. This is a matter that has been considered in some detail by Jennifer Mason (1996). This. Second. but in each a decision must be made as to what will count as representative cases. or what. a decision must be made in respect of who. a sample designed to provide a detailed. but not representing it directly. Though such samples are not impossible in interpretive research. loca- tion language. wherever possible. cases or exam- ples. however. Questions of generali- zation are tied to those of sampling because the sample is the bearer of those characteristics that it is wished to infer to a wider population. depend on recognizing the limits to generalization. Their compatibility does. processes. Her fourth type of sample is that based upon a relevant range of units. INTERPRETATION AND ANALYSIS by attempts to generalize. When this is not possible they will construct a non- probability sample (such as a quota sample) that emulates as accurately as possible a probability sample (Fink 1995: 19). or unit. will con- struct a probability sample. The case. The range referred to here might incorporate a range of experiences. close-up or meticu- lous view of cases. and so on (Mason 1996: 92). conversation etc. related to a ‘wider universe’. or . document. ad-hoc samples. statistically representative samples (generally not possible or suitable in interpretivist research). Mason (1996: 92–93) identifies four different sampling strategies in research. Sampling Once a research question is identified. in the categories and representativeness of empirical generalizations and the wider theoretical inferences that might be made from such data. Such a sample never is truly ad hoc. where no selec- tion categories are specified. Yet if it is accepted that generalization from interpretive data is a legitimate goal. types. is often seen as the only alternative to the first strategy and advises against this approach because it limits the ‘analytic potential of the study’ (1996: 92). will be the focus of that investigation. The limits to which I refer are quite separate to the objections raised by Denzin (1983) and Guba and Lincoln (1982) and arise from problems in sam- pling. The last two strategies may well be the best available. then presumably interpretivists need a sample which will reflect the relevant charac- teristics of the wider group to which they wish to generalize. First. In this section of the chapter I will discuss these limits and some of the attempts to address them before briefly setting out the case for moderatum generalizations. here may be a single individual. because the researcher must inevitably use some criteria of selection. It is however a position which rejects the suggestion that interpretations and explanation-generalization are logically incompati- ble. Third. characteristics. categories.

is unjustified (Robinson 1951: 817–818).6 This different ‘logic’ was perhaps most clearly expressed in Znaniecki’s (1934) claims for analytic induction. must decide the population about which inferences are to be made and this and the inferences themselves may not be right. According to the latter. Analytic induc- tion studies only those cases in which the phenomenon occurs. or again. the method of survey research. as defined previously. Analytic induction was developed further by Cressey (1950) and Lindesmith (1968 [1947]). What the method actually leads us to is the identification of the necessary conditions that must have been in place for the phenomenon to occur. but at least the existence (positive or negative) or non- existence of a relationship can be shown. Disconfirming cases lead to a revision of the hypothesis and the same pro- cedure is followed until further disconfirmation. enu- merative induction (statistical probability) and maintains that the former is the true method of science (1934: 236–237). until nothing new can be said. which somewhat ironically makes generalizing claims very much stronger than most in survey research. may hypothesize that there is a relationship between class and educational attain- ment. No probability can be attached to any characteristic confirming or disconfirming the interpretivist’s hypothesis. but these versions (and later ones) still do not overcome the . It is also (he says) the superior method because it allows us to establish causal relationships (rather than probabilistic ones) by identifying the essential characteristics of the pheno- mena under study. or experimenter. the assumptions made about the ‘sample’ then obtained do not hold and this was noted as early as 1951 by Robinson. Of course both the interpretivist and the survey researcher. for example. Znaniecki’s claim to be able to establish causal relationships by identifying essential characteristics. The sampling problem. The survey researcher. For example unemployment might be a necessary condition to bring about poverty in identified cases. while enumerative induction. but at least the survey researcher knows probabilistically how widely the identified sam- ple characteristics can be inferred to the population (this of course assumes an adequate sampling frame). Znaniecki’s method of analytic induction implicitly side-steps the sam- pling issue by first testing a hypothesis on a single or limited range of cases and continuing to do this until a disconfirming case is found. studies both those where it does and where it does not occur. but by studying only cases of unemployment it cannot be identified as a sufficient condition because we do not know whether it is present in the unexamined cases. Robinson maintains. For many interpretivists the last point constitutes a logical mistake – as Mason (1996: 91) claims. ‘qualitative research uses a different analytical logic’. How- ever. Znaniecki contrasts analytic induction with. or until the point is reached when further study will tell us nothing new (1934: 249). The hypothesis might have been wrong and it might have been the wrong hypothesis. analytic induction cannot discover the sufficient conditions for the pheno- menon to be explained. is indeed avoided because every case selected is part of the sample until found to be otherwise. Therefore. what he calls. GENERALIZATION IN INTERPRETIVE RESEARCH 133 relevant units.

than those which attach to people. or in sampling in interpretive research more generally. 2000 for a discussion of this) must still encounter a variant of the inductive problem. but one has to be able to recognize that cultural consistency. is a problem of decisions about categories. but they also have physical characteristics that are less transient. how do we measure that validity? Is a generalization about a fruit machine in Devon twice as likely to be right than one about a gambling strategy – or three times as likely. physical settings. Third. for example. even to non-aficionados. Second. while we can talk of (for example) generalizations about parts of the physical environment having a greater validity than those about cultural characteristics. Fisher (1993: 454). are then quite different to each other. . but they will have a validity greater than other kinds of categories that might derive from cultural characteristics. Likewise the physical description of the amusement arcade in which she made her observations. but also an approximately similar environment will enable and con- strain arcade users in other places. for example. Thus not only will these characteristics constrain or enable their users in particular ways. Normative). Categories The identification of the essential characteristics. Doxastic. Intentional. that of the ontological status of categories. tells us that most fruit machines have the same basic design. and therefore more easily verifiable in other instances. Generalizations made to other places will be of the moderatum kind. Cultural consistency may be the basis for generalization. is a familiar one. Fruit machines and gaming arcades are socially constructed. either in choosing or reject- ing cases in analytic induction. the non-equivalence of categories in respect of their relationship to each other. as Martín Sánchez-Jankowski describes in Chapter 6 in this volume. The researcher’s ability to interpret sense data and to avoid reading too much or not enough into a situation will depend on personal experiences. or equal? The problem is that the meanings of the actors can be of very different kinds. cultural artefacts. INTERPRETATION AND ANALYSIS inductive problem of inferring from a small number of known cases to those which are unknown. The categories come in many types: people. the success of representing any setting will be strongly influenced by the resources available to the researcher. The category problem takes two forms: first. language.134 GENERALIZATION. Linguistic. and second. The nature of a generalizing claim about intentional meaning (of the kind made by Bennett 1999) or of normative meaning (of the kind made by Fisher 1993) – what norms the behaviour of the individual reflects or embodies. Moreover the researcher may rely on these empirical characteristics to make theoreti- cal claims (Hammersley 1992). Generalizations have very different ontological statuses. Other equally rigorous approaches based on grounded theory (Glaser and Strauss 1967) or case study research (see Gomm et al. lists five different uses of the term (Perceptual. Kincaid (1996: 192). Symbolic.

(Hammersley 1992: 91) Hammersley is sceptical of all three justifications. as Clive Seale (1999: 109–110) points out. The third justification is that of the kind employed in analytic induction and amounts to (what I described above) as a ‘total generalization’. they are making ‘theoretical inferences’. He dismisses the first justification on the grounds that there is nothing to help the reader to decide what is of value in the situation. but it may be more difficult for anthropologists working in non-western cultures. need ultimately to reside in further empirical confirmation anyway. (c) the argument that by studying critical cases we can. but I believe that the second may be permissible under cer- tain circumstances. but that one needs experiential knowledge to know what to look for. draw inferences about the truth or falsity of universal laws. who begin their research with a lot of tacit knowledge. (b) the idea that theories are universal claims that can be derived from the study of a single case which exemplifies a type. that is they draw conclusions from their data about the necessary relationships that exist among categories of phenomena. otherwise this implies a relativism about findings that puts them on a par with fiction. GENERALIZATION IN INTERPRETIVE RESEARCH 135 In this respect the relationship of the researcher to the researched becomes important. I think he is right about the first and third. As Jankowski points out. on the basis of the hypothetico-deductive method. Hammersley describes three types of justification for theo- retical inference: (a) the claim that ethnographic work produces theoretical insights whose validity and value are to be judged by the reader.g. note tak- ing or ability to ask the right questions) and knowing something of the set- ting and the people in it. or gaming arcades in general. this doesn’t mean that only insider accounts are admissible. This might be a mixture of professional skills (e. we have to know that the particular instances were reliably represented. or on what basis they might do this. youth music. As Todd Jones (1998) asks (in respect of Geertz’s interpretation of the Balinese Cockfight) how do we know that the interpretations made are the right ones. Such claims. Theoretical inference Interpretivists often maintain that rather than making empirical generaliza- tions. The latter is perhaps less of a problem for sociologists such as Bennett (1999) and Fisher (1993). Hammersley claims that the second type of justification (b) . This view is rarely upheld in social science these days and increasingly less so in the natural sciences. what they will find insightful. that is that the analysis of individual cases can lead to the discovery of a general ‘law’. what alternatives might be available? In other words before we can say anything of Balinese cockfights.

(Hammersley 1992: 2) Ideal types are logically equivalent to theories and posit an ideal type of phenomenon (similar to an ‘ideal gas’ in physics). rather mysterious sense there are ideal types of phenomena that provide the basis for understanding any of their instances. or certain types of rational action. Prominent in this view is Clifford Geertz (1971. 1984). the result of a dissatisfaction with a form of investigation which sought to establish an authentic account (of the kind found in the ‘classical’ anthropology of Radcliffe-Brown or Malinowski). but she cannot deduce that this would be the case elsewhere. it is true to say that evidence for the existence of characteristics conforming to the ideal type in one case cannot lead us to any kind of deductive statement about the wider existence of such types and necessary relationships cannot be claimed. The change from the first to the second. Fisher (1993). in much the same way as any theory is tested. During the 1980s and 1990s there has been a shift from a hermeneutic form of inquiry. pres- sure and volume. though in general terms a consequence of the aforementioned ‘anti-positivist’ move- ment was. or will not. Though it is unlikely that the phenomenon itself will entirely conform to the ideal as described it either will. knowing that no naturally occurring gas will obey all of those laws at the same time. however much these may deviate from type. to interpretation as text centred (Denzin et al. specifically in anthropology. INTERPRETATION AND ANALYSIS assumes that in some. P O S S I B I L I T Y O F MODERATUM G E N E R A L I Z A T I O N S The ontological justification for moderatum generalizations lies with the nature of interpretation itself. which sought to uncover underlying meanings. in spirit somewhat closer to Weber and (despite the limitations of analytic induction already described) Blumer and the Chicago School than the forms of interpretive research that have evolved more recently. allowing the socio- logist to construct a picture of reality as interpreted by agents themselves. but a series of interpretations to be read as a text. or disconfirmed in this instance. ideal types of laws. 1998: 19–22). might conclude from her observations that certain types of gambling behaviour are confirmed. Analogously a physicist can posit a hypothetical gas that obeys laws of temperature. Seminal in this movement . but its influence has gone far beyond anthropology and reached its apotheosis in the fragmenta- tion of postmodernism. There seems to be no reason why ideal types cannot be proposed and evidence for the existence of characteristics resembling those of the ideal type sought. can be formulated with clarity (Weber 1978: 23) and they can be turned into testable proposi- tions. However. have features resembling the type. In the latter there is no reality to be uncovered. a story or a travelogue. Indeed as Weber noted.136 GENERALIZATION. The move toward ‘reflexivity’ anthropology is an instance of a wider ‘lin- guistic turn’ in the humanities and social sciences. for exam- ple.

then to know a culture one must come to know the language of that culture (1990 [1958]: 40–65). Such knowledge must necessarily be partial for the out- sider and filtered through his or her own perspective. Second. Moreover. first. For him rule following was linguistic prac- tice and as language is subject to variation (and shapes culture accordingly). as well as viable comparisons between places. external interpretations will be mediated through the norms and values of the interpreter. GENERALIZATION IN INTERPRETIVE RESEARCH 137 was Peter Winch’s Wittgensteinian influenced work of the late 1950s and early 1960s (Winch 1990 [1958]). to first. and second. which in its turn is more dependable than (say) statements about ‘trust regarding money’ among the players (Fisher 1993: 459). His deconstructionism. but such cultural incommensurability and distortion is surely much less likely in a British gaming arcade or music club! That is not to claim cultural homogeneity. though like Winch often an unwitting one. much of which is mediated through language. though emphasizing the separation of text as created from its creator. ignores an important characteristic of the social world. This. however. has been Jacques Derrida (1978). A further influence. Like Chinese whispers. For example. it is possible to differentiate between the validity of statements. the work of Winch. have been mobi- lized to justify a relativist interpretivism. Let us concede that cultural incommensurability may well have prevailed in the Trobriand islands. cultural variation and its linguistic expression/ motivation leads to an inherent indeterminateness which in turn produces too much variability to allow the possibility of generalization from a specific situation to others. but rather that the existence of some shared norms.8 INTERPRETATIONS AND REALITY To claim the existence of cultural consistency commits one to at least a minimal form of realism of the kind advocated by Blumer. even within settings.7 It may also be true that Geertz’s (1971) account of the Balinese Cockfight suffered from the same problems of perspective (though he par- tially recognizes this). when Malinowski did his fieldwork in the early part of the twentieth century. In other words dif- ficulty has been rendered as impossibility. the external norms and values of the interpreter render impossible any reciprocity of perspective between the interpreter and the interpreted and consequently any objective account. It is an unwarranted move from what is true that. whereby our accounts of . has been (mis)read as advocating a free play of subjective interpretation. a common language and physical referents can allow at least some reciprocity of perspective between researcher and researched. those of the kind Fisher makes about the design of fruit machines are more dependable than a statement about the players’ strategies. there is cultural variation. Moreover his accounts may be so distorted by his own positioning as to be of little value as an accurate account of life in the islands. Derrida and other thinkers such as Charles Taylor and Jean François Lyotard. its cultural consistency.

That the accounts produced are more than a ‘story’ or a ‘text’ is verifiable by those investigated. When they do not. but it is more than this. yet incomplete and erroneous (Hammersley 1989: 125). Consequently even diversity of identity creation will have its antecedents in structural homogeneity. but merely a weak inductive confirmation of one’s hypothesis. or ignored. This I believe is something Weber recognized. the cultural consistency necessary for agents to go about their daily lives. . However. the complexity of these structures and the possibility of agency to transform them. Even if the products of the fashion or music industry are more diverse than ever. but could not exist at all without a cata- logue of pre-existing cultural icons for the club goers to call upon to create their identities.138 GENERALIZATION. Those ‘realities’ as experienced are often the outcomes of processes. CONCLUSION: TOWARD METHODOLOGICAL PLURALISM In this chapter I have tried to show that generalization is both necessary and inevitable in interpretive research. they can also form the basis of testable social scientific evidence. However success in this quest cannot be seen as a firm deduc- tion. When either is present.9 The heterogeneous identities of Bennett’s (1999) youth music culture may seem fragmented and point to difference and individuality. but instead the means by which we are able to be social. Though there may be evidence of a shared reality as experienced. or shared underlying structures. this may be evidence of other mediating structural factors. in further situations. in his plea for adequacy at the level of cause and meaning. It is a denial that has its origins in the anti-positivist revolution and the consequent abandonment of hermeneutics for the linguistic turn and text-centred approaches. means that generalizations can be only moderate ones. the evidence of structures existing beyond the individuals investigated. We are attempting to describe the reality of the people we investigate. then one can hypothesize that the existence of such structures in a further situation will lead to at least some similar characteristics. There is then nothing special about moderatum generalizations and they are not just the preserve of the social scientist. INTERPRETATION AND ANALYSIS the world are regarded as substantially reliable. Indeed generalization is commonplace in interpretive research and denials of its possibility arise from a misunderstanding of what can be meant by the term. it is inadequate as a basis for policy action and for claims about what the wider social world is like. then it surely incumbent upon the researcher to produce the best evidence possible. These denied. If one is producing research which will lead to action that will change the world. the ownership and control of the production are not. Moderatum generalizations arise from that cultural consis- tency and are the basis of inductive reasoning in the lifeworld. then it is reasonable to seek the other. In generalizing from one con- text to another we can carry with us hypothetical notions of structure or out- come. Without it interpretivism is art and while art is a laudable activity. If characteristics point to particular structures in one situation.

but both as forms of scientific inquiry. For some the former term is often taken to mean all of those approaches in the human sciences that do not take a hypothetico-deductive approach to investiga- tion. The interpretations one makes of any given situa- tion have an ideographic character. While Weber did not go on to recommend methodological pluralism. whereby the weaknesses and limitations of surveys and experiments are the strengths and possibilities of interpretive approaches. such as category inequivalence and lack of repre- sentativeness in sampling.11 The view that sociology (or any social science) can be only ideographic would therefore be antipathetic to Weber (and of course the opposite would be true). These problems cannot be transcended from within interpretivism and generaliza- tions that go beyond the moderate are objectively unjustified.14 The converse is of course true. Moreover he expressed the view that with the exception of pure mechanics and certain forms of historical inquiry all ‘science’ requires each mode of inquiry (Weber 1975: 57–58). Indeed Weber (1975) speaks of nomothetic and ideographic approaches. one needs both the richness of interpretation and the ability to move beyond this to make claims about processes and structures. Their limits lie in the logical problem of inductive inference and in the ontological problem of categorical equivalence. this seems to be an inevitable conclusion if we accept that sociology has a nomothetic and an ideographic dimension. We can let them stand (and in some instances this may be sufficient)13 or we will want to develop them further. Nevertheless . while others maintain more narrowly that qualitative research is itself charac- terized by an interpretive approach (Denzin and Lincoln 1994: 2). GENERALIZATION IN INTERPRETIVE RESEARCH 139 the social world is such that in social science. the bridge between the ideographic and the nomothetic. It is these which we pick out. but also sharp features. or (more likely) as a result of some previous informal conceptual schema or more formally held theory. of specific instances. Such generalizations can be only mode- rate. a picture that has not only blurred edges. Somewhat ironically we end up with something like the working hypotheses of Guba and Lincoln (1982)12 (suggested in the quote on page 130). NOTES 1 Interpretivism and qualitative research are sometimes used interchangeably. but of course these are generalizations and as such we have to do something with them. This development implies other methods (such as the survey or experiment) that can specifically address the weak- nesses of interpretivism. They can provide testable evidence of structure and outcomes of structure. it is not so different from what a biologist or a geo- logist might do. but need only to be so. The nomothetic Weber equates with the abstract generaliz- able law like statements.10 whereas the ideographic he regards as the science of a concrete reality. The moderatum generalizations I advocate are. either as a result of striking characteristics we had not anticipated. not as scientific versus non-scientific modes of inquiry. then. Described in this way.

which might be logically expressed in the same way as law like statements. but usually the term is taken to refer to a principle of nature that holds in all cir- cumstances covered by the wording of the law. or possibly a new category between the first and second. 8 The difference between a concept of ‘strategy’ and that of ‘trust’ might be that the first could be treated as an observable in the context of the arcades. Its authenticity and reliability is not discussed. however in this case the generalizations implied would be subsumed under my second category. or how one should go about it (Kincaid 1996: 205–210). in his travelogue of Oceania. their route to so doing is littered with misconceptions about generalization (Lincoln and . Neither qualitative research or interpretivism are precise or agreed terms. for a realist account of ethnography. Certain statistical laws would allow for case exceptions and may be possible in social science (see Kincaid 1996: 63–90). There is no collective term then available to describe positivists. seeing them all as deterministic (see for example Altheide and Johnson 1998). Likewise those who regard themselves as interpretivists often differ in what counts as an interpretation. scientific pragmatists etc. Interpretive methodologists often conflate these different kinds of laws. but of course in these approaches to research generali- zation is advocated. 12 Indeed Lincoln and Guba though denying the possibility of generalization in their 1982 work later come close to adopting this position. 3 The appropriation of the term naturalism by interpretivists leaves a linguistic lacuna for the latter when they wish to describe social scientists who believe an ontological natural continuum is capable of methodological translation into scientific method to study the social world (the original use of the term). critical rational- ists. 10 Strictly speaking the term nomothetic refers to the proposal or prescription of a law. but this is too harsh and a relic of positivistic determinism. 4 Kincheloe and McLaren (1994: 151) express a similar view referring to the ‘tra- ditional’ view of external validity as ‘The ability to make pristine generalizations from one research study to another’. 6 It is not clear what is meant by a ‘different logic’ here. 7 Paul Theroux (1992). Malinowski got some things right and some wrong (Theroux 1992: Chapter 6). but I take this to mean that there is an emphasis on inductive. Thus everyone who isn’t an interpretivist must be described as a ‘positivist’ . suggesting that in generalizing. 2 What I have called ‘second-order summaries’ is evidence at third hand. 9 See Sam Porter. however.140 GENERALIZATION. rather than deductive logic. Nevertheless his observations about the greeting of strangers and the dislike of obesity were found by Theroux (who is not a trained anthropologist) over half a century later. I believe my argument here to be largely compatible with this view. It is more use- fully employed (as I do here) to mean the proposal or prescription of regulari- ties. 5 It is possible to specify a number of different kinds of 'law' in natural science. Chapter 2 in this volume. 11 More recently this point has been made by Bruno Latour (1987) who suggests that quantification is just one manifestation of providing comparative and coher- ent descriptions in science. realists. whereas the second is probably more likely to be known linguistically. They believed he had taken particular aspects of uninhibi- ted celebration in the annual Yam Festival and generalized this as typical behavi- our. found that the inhabitants knew of Malinowski. INTERPRETATION AND ANALYSIS even within survey or experimental research one can speak of variables having a qualitative dimension.

REFERENCES Altheide. Denzin. Y. (1978) Writing and Difference. American Sociological Review 15: 738–743. T. GENERALIZATION IN INTERPRETIVE RESEARCH 141 Guba 1985). Bulmer (ed. Morgan (ed. S. in May. After Schwartz and Ogilvy they conclude that perfect samples or ‘blurred’ information need not be obstacles to what they term ‘holographic’ gen- eralizations (they use the metaphor of the holograph to suggest that information about an object of study and the means to clarify it are each contained in unclar- ified versions). However if the programme is to be generalized then so must the evaluation. nevertheless they seem here to be mov- ing toward accepting the possibility of moderatum generalization. (eds) (1994) Handbook of Qualitative Research. Fielding. CA: Sage. CA: Sage.M. in N. Buckingham: Open University Press. A. Y. A. translated by A. may not imply generalization at all (Seale 1999: 107). D. (1983) ‘Interpretive interactionism’. London: Sage. Beverly Hills. (1998) ‘Quantitative and qualitative research strategies’. Derrida.. (1991) Searching for Certainty: What Scientists Can Know about the Future. (eds) (1998) Collecting and interpreting qualitative materials. Thousand Oaks. Thousand Oaks. McIntyre (eds) Readings in the Philosophy of Social Science.) Social Research Ethics. Lincoln (eds) Collecting and Interpreting Qualitative Materials. in M. New York: Basic Books. Y. Bryman. Fisher. (1999) ‘Subcultures or neo-tribes? Rethinking the relationship between youth. Geertz. London: Abacus. Bishop. London: Sage. in M. A. Fink. (1950) ‘Criminal violation of financial trust’. Cambridge. (1971) The Interpretation of Cultures. (1975) Discrete Multivariate Analysis. N. such as an evaluation of a particular programme. N. (1995) How to Sample in Surveys. Bass. Denzin and Y. London: Routledge. (1984) ‘Thick description: toward an interpretive theory of culture’. and Lincoln. style and musical taste’. P. Bryman. Of course ‘deterministic generalization’ has not been claimed. J. and Holland. 13 Of course some studies. . (1998) ‘Criteria for assessing interpretive validity in qualitative research’. Cambridge. N. S. (1993) ‘The pull of the fruit machine: a sociological typology of young players’. Sociology 33(3): 599–617. Sociological Review 41(3): 41. C. MA: MIT Press. Bennett. Casti. (1988) Quantity and Quality in Social Research. J. London: Macmillan. and Williams. (1982) ‘Observational research on the National Front’. D. and Lincoln. Denzin.) Beyond Method: Strategies for Social Research. Geertz. and Johnson. M. Fineburg. MA: MIT Press. Denzin. A. in G. N. 14 See for example Wainwright and Forbes (2000) for a discussion of the possibility of a methodological ‘third way’ in nursing research. Martin and L. J. London: Routledge and Kegan Paul. even by survey researchers for much of the twentieth century and for something to be a probabilistic generalization the odds of the sample characteristic being found in the population must be known. (eds) Knowing the Social World. CA: Sage. Cressey. C.

CA: Sage. A. R. Cambridge: Cambridge University Press. Education Communication and Technology Journal 30: 233–252. (1998) ‘Interpretive social science and the “Native’s” point of view’. Schutz. Shields (ed. Weber. London: Sage. (1994) ‘Competing paradigms in qualitative research’. and Ogilvy. CA: Sage. Milton Keynes: Open University Press. (2000) ‘Philosophical problems with social research on health inequalities’. and Lincoln. Memlo Park CA: SRE International. Denzin and Y. CA: Sage. Y. Philosophy of the Social Sciences 28(1): 32–68. (1951) ‘The logical structure of analytic induction’. Beverly Hills. London: Sage. London: Sage. (1989) The Dilemma of Qualitative Method: Herbert Blumer and the Chicago Tradition. P. Y. S. Cambridge: Cambridge University Press. Cambridge: Polity. Thornton. Lincoln. Shields. American Sociological Review 16: 812–818. M.. (1967) The Discovery of Grounded Theory. Thousand Oaks. R. (1998) ‘Influencing the policy process with qualitative research’. Holdaway. in N. R. Harmondsworth: Penguin. in W. Y. New York: Free Press. Lindesmith. (1979) The emergent paradigm: changing patterns of thought and belief. Maffesoli. (1996) Qualitative Researching. Gomm. (1992) The Happy Isles of Oceania. Latour. M.) Weber: Selections in Translation. Theroux. W. and Strauss. London: Sage. (1975) Roscher and Knies. Jones. Guba. London: Sage. (1978) ‘The nature of social action’. Lincoln (eds) Handbook of Qualitative Research. S. (1996) Philosophical Foundations of the Social Sciences: Analyzing Controversies in Social Research. E. P. Hammersley. (1988) Ethnostatistics: Qualitative Foundations for Quantitative Research. INTERPRETATION AND ANALYSIS Gephart. (1987) Science in Action. A. in M. Guba. C. A. and McLaren. and Lincoln. Mason. (1992) What’s Wrong with Ethnography? London: Routledge. London: Heinemann.M. M. P. London: Routledge.142 GENERALIZATION.) Lifestyle Shopping. (1996) The Time of the Tribes: The Decline of Individualism in Mass Society. Glaser. (1985) Naturalistic Inquiry. B. . Hammersley. and Lee. Chicago: Aldine. (1999) The Quality of Qualitative Research. E. Newbury Park. Wainwright. Health Care Analysis 8(3): 259–277. in R. Lincoln (eds) Collecting and Interpreting Qualitative Materials. Rist. (2000) Case Study Method: Divergent Interpretations. and Guba. (1982) ‘Epistemological and methodological bases of natu- ralistic enquiry’. (1968 [1947]) Addiction and Opiates. T. Runciman (ed. M. Seale. (1995) Club Cultures: Music. in N. J. M. CA: Sage. London: Sage. and Foster. (1972 [1932]) The Phenomenology of the Social World. A. in N. H. S. London: Routledge. (eds) (1993) Researching Sensitive Topics. London: Macmillan. Bulmer (ed. Kincaid. Robinson. Kincheloe. C. Renzetti. Chicago: Aldine. J. (1992) ‘Spaces for the subject of consumption’. Schwartz. R. B. Denzin and S. (1994) ‘Rethinking critical theory and qualitative research’. P. J. and Forbes. Media and Subcultural Capital. Lincoln (eds) Handbook of Qualitative Research. M. E. Thousand Oaks. Denzin and S. Weber. Hammersley.) Social Research Ethics. (1982) ‘An inside job: a case study of covert research on the police’. R.

T. (1996) Introduction to Philosophy of Social Research. Winch. New York: Farrar and Rinehart. and May. London: Routledge. Williams. F. 2nd edn. . M. GENERALIZATION IN INTERPRETIVE RESEARCH 143 Williams. (1934) The Method of Sociology. (1990 [1958]) The Idea of a Social Science and its Relation to Philosophy. M. London: Routledge. Znaniecki. (2000) Science and Social Science: An Introduction. London: Routledge. P.

Endemic within this approach was the act of representation because the societies and sub-societies studied had to be for- mally represented to both other researchers and the general public. or as part of the post- modern effort to deconstruct abstract categories by constructing alternative abstractions (Tyler 1986). and some of the cues that should be utilized to determine the level of internal and external validity present in the study. the methodological problems that such tensions provide for the researcher in carrying out their study. identifies some of the tensions that exist for the researcher when dealing with representation and responsibility. the subsequent problems associ- ated with the evaluation of evidence emerging from these studies. Thus. For example. or sub- societies. How this was to be done was left in the hands of the researcher(s) who were directly on the scene. there were generally few opportunities to replicate the study with the same group of people. This chapter . RESPONSIBILITY AND RELIABILITY IN PARTICIPANT-OBSERVATION Martín Sánchez-Jankowski Sociologists and anthropologists have had a long and prominent history of using participation-observation methodology in addressing a full range of research questions. it is about research that utilizes participant-observation methodology. I have avoided the word ethnography because the word has come to engulf a number of different kinds of studies. many of which have not done and are not interested in using participant-observation methodo- logy (Gans 1999). in a very quick and skeletal way. This chapter. researchers using participant-observation as a method were entrusted with a consider- able amount of responsibility. they were observing. Therefore. I want to make clear that this chapter is not about ethnography. The primary concern of most of the participant-observation studies was to document the everyday activities of the societies. Herbert Gans (1999) has observed that many researchers have utilized the word ‘ethnography’ to include a variety of conceptual and methodological approaches.6 REPRESENTATION. Given that a significant amount of money and effort was required to undertake this research. or as a political activity (Marcus 1988). some position ethnography as an expression of writing (Clifford and Marcus 1986). Before continuing. there is not a participant-observation study that has not involved the concepts of ‘representation’ and ‘responsi- bility’.

i. It is pre- cisely this ability that has often allowed ethnographers to believe that inter- nal and external validity are integrated into the method of data collection itself. and participating in the activities of the subjects under investigation. OBSERVATION. This filtering system will determine not simply what will be seen and not seen. ears and smell associated with the sensory system of the body. REPRESENTATION AND VALIDITY Participation-observation has the advantage of being able to directly observe the behaviour of those who the researcher is interested in studying. and the filtering system employed to gather the data. For any person (and in particular the subjects of research) it is the development of a given list of appropriate and necessary behaviours for various situations that provides the methodological problem for participant-observation researchers because .e. REPRESENTATION IN PARTICIPANT-OBSERVATION 145 on the other hand is about the methodology that has data gathered by the researcher being present. A field researcher observes and then records what he or she sees. this apparent advantage is not something that can be automati- cally taken for granted because ‘face validity’ is dependent on a number of contingent factors related to the observation of phenomena and its represen- tation. However. but how it will be seen. This knowledge-bank is of course the information and the apriority conceptualized schema that individuals have accumulated and developed to make sense of the constant stream of everyday information that must be quickly (sometimes instantly) arranged in a manner that allows a field researcher (actually any person) to decide what the most appropriate and/or necessary behaviours for a particular situation are.1 That is to say. Just how the filtering system will represent the observed phenomena will depend on the content of the system’s appa- ratus itself. In observing a certain object (a setting. I will be addressing the broad question of how observed events will be represented? The relationship between observation and representation seems to be quite straightforward. Yet the simple sequence of observation and recording are filled with critical conceptual and analytic junctures that require close scrutiny on the part of the researcher. directly observing them and the other social pheno- mena relevant to the research question. thereby representing a particular object of study. In this regard. giving the participant-observer ‘face validity’. the data analysis is occurring at the same time that the data gathering is taking place. person. The first of these critical junctures has to do with the interaction of the observer and the observed. It is the relationship between observation and representation that I first want to turn my attention to. For all participant-observers the filtering system is composed of the eyes. participant-observation. catalogue and cate- gorize. It is this directness that is the basis for the claim that participant-observation has ‘face validity’ built into it. as well as the knowledge-bank that is used to identify. thing) there exists certain points where interaction forms what I have called a ‘crit- ical conceptual and analytic juncture’.

but rather to draw attention to the fact that researchers who study a group in which they have .146 GENERALIZATION. What I am suggesting is that just because a scholar is trained and has familiarity with what the literature says about a particular area of study. The fact-value issue is one that confronts all social research. Thus. and it is not my intent to argue that schol- ars can study only the group from which they have come. However. but also must be varied. Therefore. and the emphasis here is on ‘may’. the researcher’s personal and field-research background is very important because the information accumulated during this process will have provided him or her with the cues to more accurately decipher what is important data to record and what is not (Jackson 1987: 107). but it presents an ever-present danger for the participant-observer because he or she is the data-gathering instrument.2 The types of experiences. be included in the field notes.e. gender and race. I am suggesting that the cues accumulated must not be of the same kind. there are no guarantees of course. The second obstacle associated with the researcher’s filtering system has to do with the researcher’s library of cues that help him or her determine what is going on in a particular social environment or interaction. there will be the tendency to miss either important data or interpret the observations they have made inaccurately.g. Without a public acknowledgement of the problem and some discussion as to the strategy employed to deal with it. participant-observers must remind them- selves while they are observing and recording the data. This is because the participant-observation researcher is the instrument for data-gathering. and should not. like any instrument (e. the findings are subject to questions concerning the probability of ‘type I and type II’ errors. increase the possibility of committing a type I or type II error. This is not to suggest that it is only possible for someone who has experienced the same environment through the same lenses can study that environment. As I have said. That is to say. However. and the concomitant cues that emerge from them. there is no guarantee that he or she will avoid committing errors of commission and/or omission. that they were aware of this methodological issue and took steps to reduce errors associated with it. as well as remind the reader once they are reporting their findings. By varied. The most basic examples are associated with class. a survey ques- tionnaire). The more cues researchers have in their library. knowledge base and conceptual propensity) are one aspect of the first critical juncture of representation in participant-observation research is not intended to minimize the issue of values in research. INTERPRETATION AND ANALYSIS they are dependent on their library of behaviours to determine what should. they must not come from simply one experience. The fact that the researcher’s technical capabilities (i. the library of cues not only must be extensive. because if they do. can be related to any number of categories. the higher the probability that they will recognize important conditions in the social environment that are vital in answering their research questions. if a person has no experiential knowledge of the environment it increases the probability that the data col- lected may. and thus the responsibility associated with gathering all relevant data is placed solely on the researcher.

knowing what it is that one is seeing is terribly important in determining whether such an event should be included in the notes and how it should be coded. That is to say. there is a higher probability for them to see and record data ‘relevant’ to the research question than those with less familiarity (Emerson et al. REPRESENTATION IN PARTICIPANT-OBSERVATION 147 spent a good deal of time with. places and events. When there are two or more independent observations and there is conflict in the content of the field observations there is the need to reconcile the conflict because determining how to represent an object is dependent on a consistent pattern. The second major obstacle is to determine who saw what. Or they may simply try to be. someone they think the researcher believes they are. I am referring to researchers who are both inter- acting. to minimize what they believe are their negative attributes and accentuate their perceived positive ones. but it does not necessarily produce an accurate representation of the substance. More specifically. Thus. 2000). on the other. and thus observing and conversing. one might assume that if the differences were constant that it would be clear that a problem resonated with how the researchers were ‘observing’ people. On the one hand.e. or avoid being. 1995). On the other hand. There is the tendency on the part of the subjects to act in a way that minimizes who they think they are. First. when and how. If the differences remain random (i. The act of representation presents three rather major obstacles for the researcher. if actors are the primary observational objects. The sequencing of observations is also a factor in how the object of study will be represented. The accuracy of such decisions will increase proportional to the amount of time spent by the researcher interacting with. are all the actors. and formally observing the subjects of study (Emerson et al. who they are in the particular situation under investigation. acting in a way that accurately represents: who they are personally. Sometimes this is associated with research projects in which there have been more than one field-researcher gathering data. or pheno- mena that presented itself in substantively different ways (Goffman 1959). This presents a dilemma as to how the object of study will be represented because what observations will take precedence and why determines the substance of representation. 1995: 108–112). in studies where the researcher has spent very little time in the field. The problem is to decide whether there is a substantive difference in observed behaviour. it presents the researcher with another . an alien in the environment. or a difference in each researcher’s filtering system (May et al. no pattern) then it is very difficult to identify the exact origin of the problem. with one or more of the subjects under investigation. or who they are as a proxy for a particular general group? This problem always presents itself in the beginning of the fieldwork because the researcher is. Here. this is the substance of participant- observation research. to varying degrees. Yet such an assumption would be premature because it is possible that each of the researchers were observing substantively different phenomena. there is a higher proba- bility that the researcher has committed a type I or type II error (Leidner 1993). and therefore have become increasingly familiar with.

the representation of the event. or the event’s representation of a larger object or condition. An answer like this will allow researchers a great deal of flexibility for unique situations. 1995). When reading these books. or most. Sanjek 1990. however. of data collection is another factor in the represen- tational process. or tape recording? Do the notes that are taken need to be reinterpreted and/or rewritten at some later point in time? Are they stored in notebooks.g. Recording the data requires that the researcher note down what one or more of those involved in the event observed occurring. classify. Emerson et al. Suppose the researcher is talking to one or more subjects. In those cases. It matters how notes are taken. there is a trail of verbal reflections on who said and did what. it presents an irrational system full of ambiguity for systematic comparability and replication. Dash 1997: 26–27). survey questionnaires). or at least unobserved by the researcher. or computer files? Obviously. the validity of the representa- tion can be increased if the researcher is using the news-media protocol of having three independent sources verify the action. but it cannot be authen- ticated. The researcher observes their present state and has confirmation from all. For example. shorthand. The substance of how this event will be represented is predicated on the obser- vations of the participants and not that of the researcher since the researcher was not present at this time. and writing up the results at a later time’ (Emerson et al. or manner. There are a variety of simple questions that have an impact on how sys- tematically comparable and replicable the data will be such as: Are notes taken by longhand. but unlike other methods (e. The question of what is the proper way to take notes is most likely to produce a vague answer like ‘whatever you find useful for both documenting what you observed. but the exact mechanism utilized has not been rationalized to the point that it is uniformed across research pro- jects and disciplines (Jackson 1987. to mention them raises the issue of whether the use of a computer file with its own structure and logic. INTERPRETATION AND ANALYSIS dilemma. There is no paucity of books describing the ‘how to’ of note taking. on tapes. code and scale data in a way that . In this instance. to address each of these questions adequately and consider their implications would require a separate chapter. of those involved in a particular event that the event indeed occurred. it is easy to be gripped by the events as they are recounted. remains an embedded feature in the researcher’s representation of the participants’ representations. ‘how does the author/researcher know that?’ This is particularly true when an event is being recounted and dia- logue is present (Kotlowitz 1991: 127. or its inexactness. who recount an event that just occurred to them. might provide a means to log. recently there have been a number of investigative journalists who have written books representing the lives of African Americans living in poor and high-crime neighbourhoods. The means. where and how. Thus. is predicated on unobserved phenomena.148 GENERALIZATION. and in listening to the subjects recount the event. 1995: 13–16). the question of the observational accuracy. but there are times when one wonders. This event will be composed of both verbal and physical behaviour.

In essence. representation of the ‘native’s point of view’ involved the interpretation of the native’s interpretation of the physical. if data (field notes) were made avail- able in a systematically uniformed way. Is this possible? Must a researcher become the subject they are researching? In Geertz’s (1983: 58) judgement this is neither required nor possible. This reincarnates the basic question of whether it is possible to interpret the words and linguistic . Within this phenomenological-oriented tradition. Schutz 1982). Wacquant 1995). REPRESENTATION IN PARTICIPANT-OBSERVATION 149 allows some rationality to pervade the note-taking process (Dohan and Sánchez-Jankowski 1998). Obviously the difference is substantively significant because there is a greater chance of a type II error in the first instance than there is in the second. the effort is to become the actor in order to go beyond empathy and establish understanding of the meanings that the actors of a given encounter gave to the encounter while it was occurring (Cicourel 1974. For example. and symbolic world (Geertz 1974. or patterned behaviour (i. participant- observers employing the ethnomethodological approach to their field of study believe not only that the researcher should be engaged in such ques- tions (subjective meaning for the subjects). but for many other investi- gators the research question requires precisely this. metaphysical. albeit interesting. to penetrate accurately the subjects’ understandings of reality. it would help to determine whether the account used in a paper. In both cases. multiple obser- vations). The first has to do with establishing. when studying divergent peoples and cultures (Geertz 1983). anarchistic data collection procedures will create ambiguous findings. This constitutes the second critical juncture for the researcher. economic and political) insight is based on a single observation. these researchers must put themselves in the minds of the subjects they are studying. which in turn will create strains in evaluating validity and reliability of representations within and across studies. but also that with sustained train- ing it is possible (Garfinkel 1967). In par- ticular. or it can be the meaning they gave to other actors’ activities. Without such a logical system. Thus. the primary effort is to understand how the subjects interpret their own and others’ motives and behaviours. In fact. there is the evaluative question of whether an author’s sociological (or anthropological.e. MEANING AND REPRESENTATION The representation of phenomena presents another challenge and that has to do with establishing the ‘meaning’ of an observed event or pattern of events. where appropriate to the research. This can take the form of determining what meaning the subjects gave to their own action. This is a problem that anthropo- logy has confronted. and continues to confront. report or monograph was based on a single observation or some longitudinal pattern (Johnson and Johnson 1990). the meaning of certain actions to the subjects observed. There are two broad aspects to the concept of ‘meaning’ that have implications for the participant-observer.

Thus. This pro- duces ‘rules’ (Goffman 1963) or ‘interpretative processes’ (Blumer 1969) through which actors navigate their social world.150 GENERALIZATION.e. INTERPRETATION AND ANALYSIS symbols of communicative behaviour as an accurate representation of an actor(s) meaning of an event(s). the ethnographic studies that emerge from a symbolic interactionist conceptual framework also attempt to represent the meanings that individuals bring to their social world (Blumer 1969. their social interactions (Morris 1962). does this invalidate the meanings that were recorded in the prior study? Does it nullify the meanings found in the present study? Regardless of the epistemological orientation. It also presents significant challenges to vali- dating the meaning of the representations. Thus. In both the ethnomethodologist and symbolic- interactionist tradition. as a participant-observer doing ethnography from a logical positivist tradition might. place and situation dependent. Representing the ‘common sense’ through which people play out their ordinariness is the product of this research. This would also include two rather important factors. society) of the subjects (Horowitz 1995). however. what the reader gets is what the researcher sees the subject(s) see (Wellman 1988). Similar to the concerns of the ethnomethodologists. the clouds of both a type I and type II error remain settled into an evaluative inversion. It is for this reason that ethnomethodology relinquished the quest for an answer to such a question and argued that social meaning is time. Therefore. If a follow-up study is not able to determine that the same meanings were in use as were found in a previous study. the empirical represen- tations are products of the researcher’s understanding of the meanings that the subjects bring to. 1998). but rather what the reader gets is what the researcher sees through the act of transfer- ence that accompanies their (the researcher’s) penetration of the subject’s subjective vision and its meaning (Emerson 1970). the researcher ‘interprets’ what he or she believes is the subject’s own interpretive meaning for specific situations (Fine 1996. the representation of everyday life becomes first and foremost interpretively descriptive. representation in this context has qualities that are both dynamic and static in nature. It is not something that is reproducible. Cicourel 1971). while the second has to do with determining the significance of the researcher (him/herself) to the subjects within their (the subjects’) social milieu. . The first being recog- nizing and interpreting the symbols through which meaning is communi- cated. Because there are so many individuals involved in the study. The irony is that what the researcher sees is not what the reader gets. the symbolic inter- actionist simply considers it a part of the social environment that must be considered when trying to understand the social world (i. each with their own meanings for the interactional situations they find them- selves. and get from. the representations generated from this research are built on the social meanings of behaviour as defined by the actors in relation to their interactions with each other and the researcher. Here. The result of these approaches is that what the reader gets is not what the researcher sees. with such a situation. instead of con- sidering the researcher as an object of bias. but rather.

the fact–value dilemma presents itself in ethical garb. Horowitz 1984. there are considerations as to whether the researcher ought to expose the faults of those he or she studies (Scheper-Hughes 1979). even though some did try to undermine the arguments using . actions that would hurt the population depicted in the study. researchers who did expose negative beliefs and behaviours about socio-economically vulnerable populations were in effect blaming the victim for their present condition (Ryan 1971). it is the researcher’s values that are standing nakedly before the reader. what is the proper way that they should be represented? It is at this juncture that the fact–value issue arises again and it does so in a way that Weber (1949) did not write on. but the question itself conjures up the issue of ‘responsibility’ and its relation to ‘representation’. Anderson 1990). or rationalize. The work of anthropologist Oscar Lewis (1961. 1964. they existed in both the formal and informal con- versations at professional conferences. There is no place where values plays a more prominent role than in the determination of what to include in the public version of a report. at one time the dominant position among many researchers studying the poor or underprivileged in a particular country was that the final report should not include negative. how- ever. but how they should be repre- sented. In the instance that I am now referring to. First. REPRESENTATION IN PARTICIPANT-OBSERVATION 151 RESPONSIBILITY AND REPRESENTATION The concerns about representation are not simply limited to how social. is it a violation of the ethics surrounding the social agreement made between researchers and subjects through the use of a formal protocol. there are three that are quite prominent. this is a key issue to all research. there will be differences of opinion concerning the proper ethical position to be assumed. or through the process of the researcher presenting him or herself as a close friend and con- fidant? Of course. This criticism was often utilized against scholars doing studies on racialized minorities (Frazier 1957. 1966. The arguments for this were centred on two separate propositions that were often working simultaneously with each other. The quality of the work was not the primary focus. No doubt. the evidence presented in the work could be used by politically sinister parties to carry out. Stated more simply: once the data have been gathered and analysed. or self-destructive beliefs and/or behaviours that were observed during the research. politi- cal and economic phenomena will be depicted. but it is particularly pre- sent in participant-observation research because of the closeness (both physically and emotionally) that must be obtained between researchers and subjects. Although the criticisms were not always issued in print. The second proposition was that even if the researcher was not blaming the victim for their present condition. For example. It is here that the issues surrounding the concept of responsibility penetrate participant-observation studies. Although many considerations are present. If they do. The first of these was that since the condition of the poor was assumed to be caused by structural problems. 1975) was a primary target for this criticism because he described many negative personal attri- butes of the poor he studied.

feelings of betrayal and outrage. Thus. when described in the report. Responsibility assumes a variety of characters in this context. Yet. One way that responsibility plays itself out is to have the researcher offer the subjects the opportunity to review what has been written about them. This fear is based on some level of guilt that the researcher broke the social contract that had been entered into with the subjects. In fact.152 GENERALIZATION. This fear is again associated with the subjects. The hope is that this will reduce the subject’s surprise. responsibility for some researchers seemed to dictate that what they publicly reported was only part of what they saw. public embarrassment and humiliation. A far more intense fear is that if the subjects do not like what was said.3 The upshot of this fear is to influence researchers to be very careful about what they report. responsibility is associated with letting the subjects . Thus. While Boelen’s approach was not valid. the use of negative evaluations of a participant-observation analysis by the subjects can have devastating effects on the validity of the entire research project. In a sense it is the fear that what they (the researchers) have seen will. so protecting against such an event is something that most researchers give some thought to. This occurred in a criticism of Whyte (1993 [1943]) and his Street Corner Society. there may well be an even greater feeling of guilt centred on the question of whether the research has exploited the subjects by merely treating them as tools for the researcher’s professional advancement (Scheper-Hughes 2000). be offensive to the subjects. She reported that most people refuted much of Whyte’s account of the area. Hence. what the broader research community confronted was partial evi- dence with very high probabilities of type I and type II errors. there is another fear faced by researchers that affects the ‘quality’ of the data that are ultimately presented. What would be necessary is for Boelen (or some other researcher) to study a sociological area that could be defined as a ‘slum’ and empirically demonstrate that Whyte’s findings about such areas were inaccurate. This strategy has the advantage of informing the subjects what the researcher has seen and thinks about the observations. This type of follow-up study cannot act as an adequate replication because the mere rejection of Whyte’s analysis on the basis of the subjects’ response to it is not valid. they would go before the public and declare that what the researcher reported was not the truth. The primary impetus for such criticisms was that the work could be detrimental to certain racial/ethnic minorities that were already the object of socio-economic and political discrimination. The full extent of this fear is that it will be offensive to subjects who have been very cooperative in letting the researcher carry out his or her work. INTERPRETATION AND ANALYSIS methodological criticisms. but its primary focus is on what the subjects think of what the researcher has observed. and elim- inate retaliation. even if this was a contributing factor to the subjects’ place in life. as well as the researcher’s career. In a follow-up study Boelen (1992) interviewed a number of the subjects that Whyte had identified in his study. Reflecting on such criticism there was a tendency to see that the responsi- ble thing to do under such circumstances was to avoid describing the nega- tive behaviour that was observed.

it is important that participant-observation research protect the subjects’ anonymity and confidentiality. The problem is what if the researcher removes some very important data that might be useful to other researchers. 1992). and by itself. Such conflict can lead only to confusing issues concerning loyalty to informants for aiding them in their research with loyalty to the data themselves. Even in those cases where the subjects say they don’t care about either. This has often been called the problem of invoking one’s values into the research process. thus what the subjects understand as occurring is not necessarily what is occurring. If they are compromised. or hear of a crime. but it is more than that. US law treats social scientists differently from journalists. both anonymity and confidential- ity must not be compromised. as it should. Whatever the specifics of the agreement. This is not to suggest that what the trained anthropologist and sociologist observe is the only truth. to satisfy the opinions of the subjects. what if. Failure to obey these laws subjects the researcher to criminal charges ranging from conspiracy to contempt of court. Now if researchers see. if they see. It goes beyond values. This would also include testifying before a grand jury or in a court of law. then there exists the potential for increased feelings of internal conflict about what the ‘proper’ position of responsibility should be. how should they behave? What is responsible behaviour? Is responsible . it is about understanding what is being seen for itself. However. REPRESENTATION IN PARTICIPANT-OBSERVATION 153 read the report in an effort to elicit their responses. but if professional training is to have authenticity. The issue of being responsible must not be trivialized. Responsibility is also an issue when the researcher is observing behav- iour that is illegal. of itself. psychological and legal risks. or they request that their names be made public in the report. at the same time the US federal laws pertaining to research (and monitored by each university’s committee for protection of human subjects) require that researchers protect their subjects from physical. the researcher makes changes that compromise the accuracy of the account? It is difficult at best for everyday people to be trained anthropolo- gists and sociologists. hear or in fact participate (not knowing what was going on with their subjects until it was in progress) in some illegal activity. This is a particularly acute problem when researchers go into a study already knowing that they will observe illegal behaviour. social scientists must obey all the laws that every other citizen is required to obey. As far as the law is concerned. The law in the United States governing researcher–subject relations in this context does not allow for researcher–subject privilege. make the appropriate revisions. Or even more important. doctors and lawyers where professional–client privilege is protected. it needs to identify the general dynamics associated with the observed situations which the average person’s everyday ‘local knowledge’ is not aware of (Emerson and Pollner 1988. and ultimately secure their approval. Therefore. It is important that researchers assume responsibility for honouring the agreement that they established with their subjects when they began the research. they are obligated to inform law enforcement officials.

If researchers have chosen their site (sample) carefully to match other like sites. place and circumstance to be generalizable. Thorne 1993). In the space provided for this chapter. he argued that . time or group that do not represent other persons. Thus. Duneier (1992) studied a cafeteria on the African American-dominated south side of Chicago. then reliability has been compromised and generalization is prob- lematic. if the researcher has chosen poorly. From this study. Halle 1984. Burawoy and Krotov 1992). Hagdagneu-Sotelo 1994). selecting a sample is necessary and the strategy employed in doing so will affect the degree of reliability in findings. it is not possible to discuss every issue that affects the ability to generalize. RELIABILITY AND REPRESENTATION The last issue that I want to address is associated with the question of relia- bility and generalization. ethnic groups (Liebow 1967. For example. Warner 1963). This means that the first line of representation has to do with selecting a group of people that represents other populations that the researcher wants to say something about. not all.154 GENERALIZATION. social class groups (Burawoy 1979. then one requirement affect- ing reliability has been satisfied. Even though some. One of the most important factors determining the strength of a study to generalize has to do with whether the sample population is an accurate rep- resentation of other populations in society. However. In essence. researchers operating under the symbolic-interaction tradition would not be wedded to the importance of reliability. execution and content of each study. there must be a match between the entity studied and the entity generalized about. INTERPRETATION AND ANALYSIS behaviour turning in the subjects or protecting them? This ethical predicament has an impact on the issues of validity and reliability (Venkatesh 1999). Studies can be of physical areas like communities (Lynd and Lynd 1929. place. Another way of putting this is that in the previous sections I discussed issues about what participation-observation studies will (representation) and should (responsibility) represent. places. When there is a mis- match it simply means that the researcher has chosen a person. genders (Thorne 1993. there would be a desire on the part of most symbolic interactionist researchers to have their findings generalizable to other populations. or some discrete category of groups of individuals in society (Goffman 1961. so I will simply concentrate on a few of the more important ones. Certainly ethno- methodologists are not concerned with it since they believe that observable social phenomena are too tied to a specific time. Not all the researchers who utilize participant-observation methodology are interested in the issues of reliability and generalization. The ability to generalize to other populations is dependent on a number of factors related to the design. and thus he/she has fallen victim to a potential type I or type II error (Duneier 1992. periods and groups that the researcher made claims about. and in this section I will discuss what they can more generally represent. Fine 1978). Horowitz 1984).

be represented as an empirical consistency (Leidner 1993). during a participant-observation session at the 1991 annual meeting of the American Sociological Association a discussant rose and addressed a member of the panel saying ‘What is the point of using ethno- graphy rather than other established methods like survey research if you can’t tell us something we didn’t already know?’ The rationale was that participant-observers are so close to those they are observing that they should be able to come up with some finding or interpretation that is unique and insightful (Suttles 1976). Her argument is that gender is developed socially and as such is more socially determined than the counter- psychoanalytic theory of being biologically determined. for the participant-observer. Since Thorne studied children who were already in school. There is a great deal of pressure from researchers utilizing other methods of social inquiry to have participant-observers find something unique and different. and they should not. Her study looks at the issue of gender formation in children. One of the most troubling aspects to evaluating participant-observation research has to do with whether the researcher is presenting a finding that has come from an established pattern in the data or simply an observation that is the most interesting. and there was an intuitive feeling within the broader research community that the argument was probably correct. The length of time that researchers spend in the field will have a direct impact on the ability to generalize their representations to other populations or situations. She studied children at play in a school playground. they had already passed the time when the counter-theory says that gender identity has been established. but they cannot. it is unlikely that a sufficient amount of data will have been accu- mulated to establish a pattern. Incomplete patterns are simply that. Thus. For example. the study itself was very vulnerable to a type I error.4 The study by Thorne (1993) is another example of a potential type I error. They can be suggestive. we are here again faced with a potential type I error. Such a sample can only tell you how they behave in a cafeteria like the one they are in. time or event observed. This is because the sample he chose was not able to reject the argument of the other contending studies. finding one interesting observation and building it into a generalization is very tempting . REPRESENTATION IN PARTICIPANT-OBSERVATION 155 the claims made in much of the other research depicting African American men as irresponsible were inaccurate. When researchers spend small amounts of time doing field- work (weeks or a few months) there is the question of whether the actions being observed are representative of some general pattern or are idiosyn- cratic to the person. The problem here was that while the argument that he was advancing was socially appealing. While this may be true. Without significant amounts of field time. Observing people’s behaviour in a cafeteria cannot tell the research how they behave (responsibly or not) in environ- ments like work or the family. there is no way to rule out the claims of the contending theory unless a longitudinal study following very young children through the ages when the psychoanalytic theory would have predicted that gender identity would occur is done. Thus.

the potential for a type I or type II error is great enough to make reliability tenuous at best. INTERPRETATION AND ANALYSIS indeed. and second. but two of the more significant ones are. Although a full elaboration of the problem can- not be done here. first. reliability becomes more tenuous and so does the ability to generalize. Often this situation is created when a team of researchers is utilized to collect the data. and not an insignificant chance. Yet. there is a problem when the primary interpreter of what the prin- cipal investigator is representing as his or her own direct observation of behaviour is in fact not that of the researcher. The final element impacting generalizations (i. but that of a key informant. with limited numbers of observations. 2000). and this in turn impedes the researcher’s effort to generalize (May et al. the princi- pal investigator has designed a comparative project and does not think that he or she can do all the sites in the time desired. generalized to. representations) has to do with the access of the researcher to those he/she is observing. because of a mismatch between his or her physical or social characteristics and those of the subjects. Ultimately a number of issues related to the investigator’s observations influence how the subject matter can be represented. Sullivan 1989). as opposed to some other methodological . Finally. in brief the problem is that there is no way to judge the inter-researcher consistency in what each independent researcher observed.e. and is usually responsible for the researcher having obtained access to the subjects (Whyte 1993). This situation produces problems in internal validity that has a direct impact on reliability. the principal investigator does not. feel that he or she can gain the closeness necessary to execute the research properly (Moore 1978. CONCLUSIONS: THE THREE R s IN PARTICIPANT-OBSERVATION One of the prevailing beliefs about participant-observation research is that it provides a clear and vivid picture of everyday life. It could very well be accu- rate. This occurs for a number of reasons. The obvious problem is that the informer is not trained in one of the social sciences and this raises a question as to whether the informer’s interpretation of behaviour is accurate. but there is also a chance. it is this ability that has often been utilized by both participant-observers and survey research- ers to justify the very use of this. compo- sition can. Whatever the reason. In fact. This assump- tion is unfounded because there is a great deal of variation in the level of closeness between researchers and those they are observing. and are.156 GENERALIZATION. when teams are used there is an increased risk that the data are unreliable. that it is not accurate. to the subjects they are studying. There are times when the researcher conducting the analysis part of a study was not the person who collected the data (Sullivan 1989). or similar. and this influences how other objects of like. There is an assumption that all participant-observers are by definition close. Here again. both physi- cally and communicatively. This informant usually has been with the researcher from the start of the study.

Like any other method. but . External validity is the consistency between the way that a particular concept or construct is defined for the study and the way that it is used in the standard literature. When both quantitative and qualitative researchers utilize such evaluative care. systematic and generalizable. this is an assumption that cannot be accepted in an unqualified way. Rather. there are a significant number of technical methodological issues that permeate the findings of each participant-observation study and many of the published participant-observation studies have serious problems that require reservation in fully accepting their findings. In this way. A type II error is defined as rejecting a research statement as being false when it is in fact true (Blalock 1972). ‘what they see as researchers is what readers get. it must be evaluated systematically. In other words. some parts are and others are not. he felt compelled to defend himself at a session of the annual meeting of the American Sociological Association and in the fourth edition of his book (Whyte 1993: 360–371). what they saw?’ NOTES 1 Internal validity is the consistency between the operationalization of a concept at two or more points in the data analysis stage of the research. REPRESENTATION IN PARTICIPANT-OBSERVATION 157 approaches. 4 There was also the issue as to whether Duneier’s choice of a cafeteria was typical of an African American eatery. and statistical techniques of analysis are the only reliable means to research a question. 3 William Foote Whyte was very upset by the Boelen (1992) piece because he thought it was a direct attack on his integrity as a researcher and a person. ethnography will move beyond the novel.’ Yet. By not taking the proper care to eliminate potential type I and type II errors in participant-observation research a significant number of these studies remain confronted with the question: ‘Is what we’ve got. There is no question that anthropo- logical and sociological ethnographers have contributed to that prejudicial argument. as well as from the failure of ethnographers to be careful in planning and executing their research. or the entire study is compromised because of some endemic methodological flaw. as I have argued in this chapter. One of the primary points that I want to make in this chapter is that an ethnographic piece of research should not be accepted or rejected merely on face value. 2 A type I error is defined as accepting a research statement as being true when it is in fact false. The cafeteria was on the south side of Chicago. an assess- ment can be made as to whether the entire study is reliable. At the present time. interesting and anecdotal and assume the ranks of the rigorous. Much of this justification is based on the belief that ethnographers simply detail life as it is. participant-observation studies are generally not considered capable of proving or disproving a study that has utilized survey data with large samples in generating a set of findings and conclusions. This prejudice comes from both survey researchers’ belief that large samples. standardized questions asked in systematic ways. Thus.

L. in H. (1997) Rosa Lee: A Mother and her Family in Urban America. L. Cicourel. Burawoy. vol. (1988) ‘On the uses of members’ reponses to researchers’ accounts’. P. and Sánchez-Jankowski. M. E.V. Chicago: University of Chicago Press. and Shaw.A. M. REFERENCES Anderson.M. Blumer. Burawoy. American Sociological Review 57: 16–38. Chicago: University of Chicago Press. Miller and J. Dohan. 3. New York: McGraw-Hill. (1990) Streetwise: Race and Urban Change.A. New York: Free Press. Fine. (1996) Kitchens: The Culture of Restaurant Work. New York: Plume. (1970) ‘Behavior in private places: sustaining definitions of reality in gynaecological examinations’.158 GENERALIZATION. Hagan (ed. the issue of gen- eralizing about how African Americans behave in a typical African American eatery is also vulnerable to a type I error.A. MA: Harvard University Press. Human Organization 47: 189–198. Dreitzel (ed. Thus. Holstein (eds) Perspectives on Social Problems: A Research Annual.A.I. CA: University of California Press. H. CT: JAI Press. A. Duneier. R. (1969) Symbolic Interaction. H. (eds) (1986) Writing Culture: The Poetics and Politics of Ethnography. Fretz.) Recent Sociology No. Englewood Cliffs. Boelen. Journal of Contemporary Ethnography 21(1): 11–51. Palo Alto. M. CA: Annual Reviews. Emerson.) Studies in Interaction. Clifford. M. Cambridge. (1971) ‘Basic and normative rules in the negotiation of status and role’.E. and Masculinity. Respectability. A. (1998) ‘Using computers to analyze ethno- graphic field data: theoetical and practical considerations’. J. M. (1978) With the Boys: Little League Baseball and Preadolescent Culture. (1992) ‘Cornerville revisited’. (1979) Manufacturing Consent: Changes in the Labor Process under Monopoly Capital. G. New York: Free Press. Greenwich. Blalock. D. R. in D. and Pollner. R.M. Fine. INTERPRETATION AND ANALYSIS that part of the south side that is adjacent to the University of Chicago. 24. (1992) Slim’s Table: Race.A.M.M. . Chicago: University of Chicago Press. (1998) Morel Tales: The Culture of Mushrooming. Emerson.P.. Dash. Cicourel. in G. (1974) Language and Meaning in Social Interaction. Emerson. NJ: Prentice-Hall. G. R. which is an area that is wealthier and more racially integrated with white professors and students than nearly all the other areas on the south side. Berkeley. Sudnow (ed. and Krotov. (1972) Social Statistics. Fine. and Pollner. G. New York: Macmillan. (1995) Writing Ethnographic Fieldnotes. Berkeley. (1992) ‘Difference and dialogue: members’ read- ings of ethnographic texts’. Chicago: University of Chicago Press.M. 2: Patterns of Communicative Behavior. Emerson. L. M. CA: University of California Press. J. Chicago: University of Chicago Press. W. vol. (1992) ‘The Soviet transition from socialism to capi- talism: worker control and economic bargaining in the wood industry’. and Marcus.V. in J. G.) Annual Review of Sociology.

Horowitz. R. O. Geertz. NJ: Princeton University Press. Horowitz. W. (1959) Presentation of Self in Everyday Life. (1987) Fieldwork. NY: Cornell University Press. (1963) Behavior in Public Places: Notes on the Social Organization of Gatherings.J. and Johnson. (1984) Honor and the American Dream: Identity and in a Mexican American Community. Chicago: University of Chicago Press. Bulletin of the American Academy of Arts and Sciences 28. CA: University of California Press. and Pattillo-McCoy. Berkeley. Marcus. IL: University of Illinois Press. Goffman. Lewis. . (1995) Teen Mothers: Citizens or Dependents? Chicago: University of Chicago Press. (1983) Local Knowledge: Further Essays in Interpretive Anthropology. O. E.) Fieldnotes: The Making of Anthropology. Qualitative Inquiry 6: 65–87. May. CA: University of California Press. (1984) America’s Working Man: Work. Berkeley. (1964) Pedro Martinez: A Mexican Peasant and his Family. E. Lewis. New York: Free Press. (1966) La Vida: A Puerto Rican Family in the Culture of Poverty – San Juan and New York. NY: Doubleday. Englewood Cliffs. New York: Anchor. (2000) ‘Do you see what I see? Examining a collaborative ethnography’. Drugs and Prisons in the Barrios of Los Angeles.M. R. D. PA: Temple University Press. NJ: Rutgers University Press. (1974) ‘“From the native’s point of view”: on the nature of anthro- pological understanding’. C. (1990) ‘Quality into quantity: on the measurement potential of ethnographic fieldnotes’. A. Moore. E.W. Journal of Contemporary Ethnography 28: 540–548.R. (1957) Black Bourgeoisie: The Rise of the New Middle Class in the United States.E. R. (1929) Middletown: A Study of Contemporary American Culture. MA: Little. A. Lynd. Princeton. J. Buford. Johnson. Morris. in R. Gans. H. O. (1967) Tally’s Corner: A Study of Negro Street Corner Men. (1999) ‘Participant observation in the era of “ethnography”’. Jackson.. M. Urbana. H. and Society From the Standpoint of a Social Behavioralist. New York: Free Press. O. New York: Random House. C. New York: Pantheon. Boston. Liebow. R. Lewis. E. (1991) There Are No Children Here: The Story of Two Boys Growing Up in the Other America. O. New York: Random House. New York: Basic Books. Lewis. and Politics among Blue-Collar Property Owners. Self. (1988) Ethnography through Thick and Thin. Home. NJ: Prentice- Hall. Geertz. Garden City. Hagdagneu-Sotelo. Goffman. Leidner. (1967) Studies in Ethnomethodology. Ryan. Goffman. (1994) Gendered Transitions: Mexican Experiences of Immigration. (1961) Children of Sanchez. REPRESENTATION IN PARTICIPANT-OBSERVATION 159 Frazier. New York: Anchor. (1978) Homeboys: Gangs. Philadelphia. Kotlowitz.S. P. (1975) Five Families: Mexican Case Studies in the Culture of Poverty. and Lynd.F. Sanjek (ed. (1962) George Herbert Mead: Mind. A. R. (1971) Blaming the Victim. E. B. Garfinkel. Chicago: University of Chicago Press. New Brunswick. C. Ithaca. (1993) Fast Food. Fast Talk: Service Work and Routinization of Everyday Life. H. New York: Random House. (1961) Asylums. G. New York: Random House. New York: Harcourt Brace. Brown. Halle.

(1999) ‘The promise of ethnographic research: the researcher’s dilemma’.A. Ithaca. (1993 [1943]) Street Corner Society: The Social Structure of an Italian Slum. Whyte. (1990) Fieldnotes: The Making of Anthropology. R. Tyler. S. M. Wacquant. Thorne. Finch. Chicago: University of Chicago Press. W.D. (1976) ‘Urban ethnography: situational and normative accounts’. NJ: Rutgers University Press. N. (1963) Yankee City. B. Berkeley. NY: Cornell University Press. Sullivan.L. S. translated and edited by E. A. Ethnography 1: 117–140.A. Wagner. Shils and H. (1986) ‘Post-modern ethnography: from document of the occult to occult document’. Clifford and G. (1949) The Methodology of the Social Sciences. M. Weber. Suttles. N. 4th edn.160 GENERALIZATION. NY: Cornell University Press. translated by H. Annual Review of Sociology 2: 1–18. (1982) Life Forms and Meaning Structures. Marcus (eds) Writing Culture: The Poetics and Politics of Ethnography. in J. Symbolic Interaction 11: 59–68. G.E. (1995) ‘The pugilistic point of view: how boxers think and feel about their trade’. D. Schutz. Ithaca. W. New Haven. (1988) ‘The politics of Herbert Blumer’s sociological method’.L. Wellman. New York: Free Press.A. CA: University of California Press. Warner.F. New Brunswick. Scholars and Schizophrenics: Mental Health in Rural Ireland. (1989) ‘Getting Paid’: Youth Crime and Work in the Inner City. (2000) ‘Ire in Ireland’. Law and Social Inquiry 24: 987–991. Venkatesh. London: Routledge. INTERPRETATION AND ANALYSIS Sanjek. Berkeley. CA: University of California Press. (1993) Gender at Play: Girls and Boys in School. L. Scheper-Hughes. (1979) Saints.R. . Scheper-Hughes. CT: Yale University Press. Theory and Society 24(4): 489–535.

7 AUTOMATING THE INEFFABLE: Qualitative software and the meaning of qualitative research Nigel G. The point will be made that patterns of use are as apt to reflect the research environ- ment as the technology. the last practice to attract such attention has been that of data . Fielding Computers have been used for half a century in research using textual data but the introduction of software to support qualitative analysis has proven controversial in a way that it never was in the field of content analysis. practical and research environment implications. WHAT IS COMPUTER ANALYSIS OF QUALITATIVE DATA? Computers have already established a considerable presence in qualitative research. data collection and literature searches have attracted little comment. We note that for most users the story of qualitative software is the story of the analytic procedure called ‘code-and-retrieve’. Epistemological preoccupations are more enduring than any technology. Yet the information technologies which now support our writing. including approaches to data analysis which are implicitly numerical or are based on formal logic. However. from data collection through analysis to presentation (Weitzman and Miles 1995). Computers are routinely used at each stage of qualitative projects. But we will also sketch in other promising applications of qualitative software. The chapter moves on to explore how people actually use qualitative soft- ware. we begin with a description of qualitative software. demystifying what sometimes seems a technical ‘black art’. But we will also observe that software could play a significant part in changing the workings of the qualitative research com- munity. To understand this controversy. in contrast to the debates prompted by the emer- gence of software to support qualitative data analysis. the epistemological implications of qualitative software do not mirror the extent of its technical. despite its self-image as a craft. and its implications for qualitative research. in a field which has been slow to document and codify its procedures. It may not be coinci- dental that.

the notable thing about qualitative software was that it was developed by social scientists. when the first packages emerged for use on mainframe computers. A dominant concern about qualitative software is that it may somehow ‘take over’ the analysis. even those indifferent to qualitative software may concede that its emergence has benefited the field by obliging us to be more explicit about how we manage. This makes for close awareness of developments with other packages and responsiveness to the needs of users. The prototypes may have suffered bugs and for- bidding interfaces but they were closely informed by what researchers regarded as essential to support qualitative data analysis. It is sometimes forgotten that the monster was a threat only because he was abused by humankind. INTERPRETATION AND ANALYSIS analysis. each displaying considerable distinctiveness. and there is no prospect that it will ever excuse the need for researchers to think. Each has merits and deficien- cies which suit it better for particular kinds of work. make the analytic process more ‘transparent’ and accountable. even now. Kelle 1995). Although com- mercial pressures inevitably intrude as the software becomes more profes- sional. so they require a means to ‘navigate’ the database. analyse and interpret qualitative data.162 GENERALIZATION. Indeed. The imposi- tion of a standard approach is obviated by the sheer variety of packages. quite limited in the kinds of support it offers for analysis. Software for computer-assisted qualitative data analysis (CAQDAS) seeks to facilitate data management chores which are tedious and subject to error when done manually. Researchers need to move around the corpus looking at the relevance of specific analytic ideas to different segments. systematic and quantitative description of the manifest content of communication’ (Berelson 1952: 489). software must be able to store data in different formats and enable researchers to annotate it so as to track the effects of field relations on it. the so-called ‘Frankenstein’s monster’ debate (Lee and Fielding 1991. and most remain social scientists as well as developers. produce ‘concordances’ (organized lists of words and phrases) and quickly access the context of particular text segments. though. There is another kind of rebuttal to the take-over fear. address text units as ‘variables’. imposing a standard approach and employing con- cealed assumptions. both practically and analytically. To marshal the variety of data that qualitative researchers use. and support analytic approaches which would otherwise be cumbersome. Qualitative software is. developers still exchange ideas via conferences and publications. Researchers need to determine which best fits the kind of work they do. Programs had to store extensive documentary data. The mode of analysis meant programs required statistical features or to allow data to be exported to statistical pack- ages. Development was often in response to data management or analysis require- ments on specific projects. not computer scientists. The earliest computer application to textual data was content analysis. These requirements produced software of limited use to qualitative researchers. which pursues ‘objective. From the late 1970s. .

Users may also do searches which recover only data where two particular characteristics apply but not a third. ‘Multiple sort’ retrievals handle cases where one category is considered in relation to another. let users recover the data pertaining to each category where keywords appear in the data. WinMAX and The Ethnograph support the division of text into segments. wherever these words appear in the text the software will extract it. including HyperQual. For exam- ple. and analytic memos can be linked to the data. Retrieved text can be sorted into new files. can be retrieved. attaching codes to the segments. ZyINDEX and Sonar Professional. A ‘single sort’ retrieves all the data pertaining to one code. or have patterns like the sequences of letters and numbers in social security records. Text retrievers are fast at retrieval. Researchers also need to be able to write ‘analytic memos’. code-and-retrieve packages and theory-building software (Weitzman and Miles 1995). If the respondent did not use the keyword it must be added to the data so the segment can be retrieved. ‘Coding’ is fundamental to qualitative data analysis. AUTOMATING THE INEFFABLE 163 They must also be able to retrace their steps in these operations. and retrieving segments by code (or combination of codes). mean the same thing. including facilities to handle quantitative data. data analysis is a process of data reduction. Researchers aim for codes which capture some essential quality of the segment. Text retrievers. A TYPOLOGY OF QUALITATIVE SOFTWARE Coding and retrieval of coded segments informs all three basic types of quali- tative software: text retrievers. you may have a category called ‘social class’. data from MALE . in one or many files. ethnomethodologists and conversation analysts need to assign to segments a term for routines which recur in behaviour/talk). The corpus has to be divided into segments and these segments assigned codes (or categories) which relate to the analytic themes being developed (similarly. Words. Otherwise every segment would be a code and we would get nowhere. either globally or in application to particular segments. Content-analysis capabilities are often pro- vided. Sophisticated variants like askSam and FolioVIEWS have more ways to organize text and make subsets. Code-and-retrieve packages. for instance. giving the reasons why they assigned a given code to a segment or part containing the germ of an analytic point they want to develop. other character strings and combinations of these. When you search for ‘social class’. WordCruncher. Kwalitan. Some deal with highly structured text organized into ‘records’ (spe- cific cases) and ‘fields’ (numerical or text information appearing for each case). When data are coded we can retrieve them selectively using various analytic strategies. Their forte is information manage- ment where basic data is held on large numbers of people. and which apply to other segments too. for example. for example where data coded ‘age’ coincides with data coded ‘social class’. along with things that sound alike. such as Metamorph. During this process we sometimes need to revise codes.

1). for example. Researchers sometimes need to retrace their interpretive work. The top left window lists the segments pertaining to the ‘tutoring style’ code. ‘tutoring style’ is highlighted as it is the code currently being considered. Code-and-retrieve packages approach the writing of memos in different ways. Note the icon in the extreme left margin. WinMAX offers an example of contemporary code-and-retrieve software (Figure 7. retrieved so the user can compare them. particularly in team research. the user clicks on the icon.1 EXAMPLE OF CODE AND RETRIEVE SOFTWARE respondents living in URBAN areas who are NOT married.164 GENERALIZATION. The bottom right window is a hierarchical list of the texts being analysed. an extract from correspondence with a prisoner about the prison education programme. INTERPRETATION AND ANALYSIS FIGURE 7. The window at top right lists codewords hierarchically. to see it. it indicates a memo has been written relating to a seg- ment of the extract. Ideally the software reminds users that an analytic memo pertains to a segment every time it is inspected and displays the memo instantly if desired. Various types of Boolean retrieval are supported. with main codes followed by sub-codes. ‘or’ and ‘not’ relations of Boolean mathematics). to check coding has been con- sistent. Hypertext features allow users to move quickly around (navigate) the database. Such strategies are called Boolean retrievals (following the ‘and’. In the next margin are markers indicating the extent of segments of the extract to which codes have been applied. At bottom left is the window containing the raw data currently being worked on. .

Yet the comparability of qualitative data is problem- atic. and so on. Some can show code names (or other objects. Full Boolean searching and the ability to ‘test’ hypotheses may be offered. to be certain there are no cases where the conditions postulated for the phenomenon to occur appear in the absence of the phenomenon (Hicks 1994 called this ‘neo-analytic induction’). on the basis that it occurs only when par- ticular conditions apply. Users hypothesize a relationship between the occurrence of a particular statement in the data and the occurrence of another (if ‘this’ then ‘that’). To construct a testable hypothesis users may need several ‘if’ and ‘then’ rules. Case-oriented approaches. Rapport with respondents may vary. Theory-building software. AUTOMATING THE INEFFABLE 165 Such software focuses analytic attention on relationships between codes (or categories) and data. Qualitative Comparative Analysis (QCA) uses logic . There is a fundamental distinction between analysis based on codes and analysis based on cases. respondents may talk with differ- ent degrees of specificity. The idea is to isolate the essential conditions deter- mining the phenomenon. such as Atlas/ti. Users have to assume that the data supporting these hypothetical relationships is comparable and that coding has been done in a commensurate fashion. and N-vivo. especially where there is an interest in causality. HyperRESEARCH. because they are working ‘one level up’. Cases are differentiated by the relationship between their outcomes and the components they share with other cases having different outcomes. Clearly such procedures take us some way from working with the data itself. A controversial example is the ‘hypothesis test feature’ in Hyper- RESEARCH. Neo-analytic induction is useful in analysing dynamic processes. In fact. Hypothesis-testing will be discussed further. NUD*IST4. this can be made part of another hypothetical relationship. although code-and-retrieve support is also offered. The approach called Qualitative Comparative Analysis (Ragin 1987) thinks in terms of outcomes. formulate propositions which fit the data and test their applicability. or visualise connections between cate- gories as an aid to conceptualization. The hypothesis test searches for cases where the particular codewords occur in the particular combinations the proposition requires. These packages help users develop higher-order classifications and cate- gories than those derived directly from data. after considering another formal approach. like memos) as nodes in a graphic display and users can link them to other nodes by specified relationships like ‘leads to’. the basis of analytic induction. emphasizes relationships between the categories. Analytic induction considers instances where some phenomenon of interest is present. Cases are literally recognized as such from the different outcomes they display. Such software can facilitate formal approaches to qualitative data analysis. ‘is a kind of’ and so on. each using codewords assigned to the data. If these are present that case supports the proposition. When the two occurrences are found together. suit some analytic purposes. Work with codes is not the only approach to qualitative data analysis. with relationships between the categories/codes. comparing outcomes across multiple cases. analytic induction also needs to consider cases where the phenomenon is not present.

unlike HyperRESEARCH. comprising elements and their connections. and its ‘intensions’ consist of all attributes whose names are attached to the node or to a node derivable by an ascending path from the node. The program enables users to model narrative accounts of event sequences and produce mathematically based causal accounts. A concept can be graphi- cally displayed in a line diagram or ‘concept lattice’ consisting of nodes. Although the data that Qualitative Comparative Analysis and Event Structure Analysis use are qualitative. INTERPRETATION AND ANALYSIS and mathematics to identify systematically ‘universal conditions’ which are always present in particular combinations when the phenomenon occurs. The descending/ascending element reflects the hierarchy or ‘levels’ of data and concept. lines and labels. They suggest that. but. it supports the iterative process of formulating codes/categories familiar from grounded theory and which lies behind the code-and-retrieve packages. Event Structure Analysis is sup- ported by a program called ‘Ethno’ (Heise 1991). Based on these premises abstract logical structures of events are generated. There is a further formal approach that should be mentioned because it strongly embraces concerns about the importance of context. Analysis proceeds by traversing these paths from objects to attributes. . if the data are going to be so stringently ‘reduced’. and certain events cannot occur before their prerequisites. The status of these formal approaches is an important issue to which I will return. This approach is Formal Concept Analysis (FCA: Mackensen and Wille 1998) and. there is little point in collecting qualita- tively ‘rich’ data in the first place. they also require extensive data reduc- tion prior to analysis. Like QCA it seeks explanatory models for interpreting sequences of events (Heise and Lewis 1988). some do not regard these approaches as qualitative analysis at all. The prototype application is based on the TOSCANA data management system. although it is mathematically based (on set theory). or by following links from items of data to categories linked at several progressive levels of abstraction. investigating the concepts by ‘zooming in’ from a more abstract level to the data. The extensions of the concept consist of all objects whose labels are attached to the node or to a node derivable by a descending path from the node. Event Structure Analysis represents series of events derived from fieldwork as logical structures. having a chronological (process) orientation and concern with causal explanation.166 GENERALIZATION. Its basic logic is that in each situation social actors have limited choices. Their concern is that the analysis proceeds at a level too far removed from the original data and the context in which they were collected. Like the hypothesis-testing approach of HyperRESEARCH. We should note a further formal analytic approach using qualitative data. FCA construes concepts as con- stituted by their ‘extension’ – all objects belonging to the concept – and ‘intension’ – all attributes valid for those objects. and compared to actual event sequences. The ‘QCA’ software analyses conjunctures of particular elements of cases which bear on the outcome of a process. Qualitative Comparative Analysis and Event Structure Analysis involve high-level cod- ing.

We can no longer argue that the software is simply an aid to code-and-retrieve or that code-and-retrieve is the essence of qualita- tive analysis. The FCA approach uses mathematical techniques of conceptual and logical scaling to produce its concept lattices. The first generation were word-processor and database management systems. As development proceeds and packages become more fully featured. have users’ practices changed accordingly? WILL QUALITATIVE SOFTWARE KILL OFF MANUAL METHODS? Many qualitative researchers remain committed to manual methods and. The question is. with the ability to search for co-occurring codes. The main analytic metaphor was code-and-retrieve. and to dependencies among categories. the main example being the Windows operating system. AUTOMATING THE INEFFABLE 167 FCA offers a means to represent conceptual systems graphically while being able to return at any time to the data. The generational perspective has an important implication. and various facets of these sources. The three-fold typology of qualitative software can be regarded as succes- sive generations (Kelle 1996). to empirical exam- ples which contradict elements of postulated concepts. it was informed by prevailing approaches to analysis. and few looked to software to inspire new approaches to analysis. Exchanges between users and developers motivated many changes. for . Mangabeira (1995) identifies the distinctive element of this third generation as being the model-building capabilities of the software. and meth- ods to construct complex networks linking categories. The pattern of diffusion of the ‘simplicity’ idea could be traced and key sources identified (as nodes with many links). a measure of convergence occurs. for example. provided facilities to support theory building and hypothesis testing. Specialized code-and-retrieve programs developed in the early 1980s con- tributed the ability to manage unstructured textual data and sophisticated cut-and-paste. others were imposed. codes. For example. which supported techniques like cut-and- paste but did not exceed the most basic requirements of data management. Most importantly. as a code-based theory builder rather than a code-and-retrieve package. manual methods have advantages over software. However. The representation alerts users to con- cepts which are thinly supported (have few links to data). packages increasingly support procedures which are new or impractical without the computer. allowing inferences about the degree of authority contemporaries accorded each source. the Windows release of the first generation package ‘MAX’. including their date of publication. indexing and memoing techniques. the second generation made coding and retrieval processes transparent. one might regard WinMAX. but is more in line with the premises of qualitative research than the other formal approaches. memos and text segments. for certain purposes. The third generation. When qualitative software first emerged. research on the emergence of the notion of ‘simplicity’ as a criterion in classical music drew on a set of published sources in which the term was used. Consequently the typology becomes less rigid.

This may relate to the hope that CAQDAS can resolve time pressure. crashes. which. and those relying on manual methods may find themselves having to invest inordinate effort in the paper-based equivalents of procedures which are done in seconds using software. a major benefit of qualitative software is that it obliges researchers to be clearer about their reasoning. the consensus is that people should have a reasonable grasp of the foundations of qualitative research first (Allatt and Benson 1991. This also avoids con- flict between learning the rudiments of method and learning the procedures of software use. which tends to exploit data management rather than conceptualizing or analytic features (Fielding and Lee 1998). for example. However. Such prompts are only as good as the criteria ‘wired into’ the pro- gram. It is true that pro- totype ‘expert systems’ have emerged. CAQDAS IN THE RESEARCH ENVIRONMENT Sales data for one of the longest established packages. Accounting for respondents’ experiences means exploring how ‘software issues’ impact on ‘research environment issues’. and so on. although such hopes appear to be false (Lee and Fielding 1995). technical and practical issues. where setting up the data in the required format would take a disproportionate time relative to the scope of the analysis. Tangled up with software use were fundamentals of the social context of research – sponsors’ demands. It is important to repeat that simply using CAQDAS does not mean the whole analytic process takes place ‘within’ the software. and enables the analytic process to be more transparent and therefore accountable. prompt users to consider whether the data with which they have supported an inference is adequate. INTERPRETATION AND ANALYSIS example. These characteristics will shift com- munity standards of analytic quality over time. One might assume that research on CAQDAS users would be dominated by reports of bugs. The Ethnograph. deadline pressures. program architecture and so on – that is. to gain a firm grounding in procedures that they may later automate. sug- gests that early users featured in applied fields such as social research in medical settings rather than in discipline-based academic research.168 GENERALIZATION. No one should feel guilty about not using software or pressured into adopting it. . the focus groups and individual interviews that Ray Lee and I conducted with about 60 CAQDAS users in Britain (Fielding and Lee 1998) revealed that the issues preoccupying users were not simply technical but matters of the ‘research environment’. Research suggests this pat- tern of adoption may also account for the typical mode of use. Fielding and Lee 1996). It is pedagogically desirable that those new to qualitative methods should have their initial experiences using manual approaches. power relations in research teams and with research sub- jects. expert system software has a potential role in teaching. but the proto- types encourage users to override program advice if they wish. and vice versa. However.

Computing centres remained unwilling to help with support and training. the organization of the software market. the seemingly straightforward matter of obtaining a package ready for use on one’s own machine invokes issues concerning the status of qualitative research. applied researchers face their own problems. even now. we found no case where users reported CAQDAS saved on analysis time. and some felt they were under- utilizing the package. in an exchange on the email list for Atlas/ti users a would-be user in Texas appealed for training. largely ad hoc and patchily distributed. Social science computing began and ended with statistics packages. and other factors (for full discussion of ‘choice’ issues. Not only have developers not claimed that using CAQDAS is . solo/team and so on). for example. We documented cases where users faced opposition to CAQDAS from senior project staff or doctoral supervisors. but because someone in their social network had details of a particular package. Training is. and research sponsors appear to favour CAQDAS for reasons of legitimation. This is probably not susceptible to a tech- nological fix. which elicited an offer if the individual could get to Santa Barbara. the balance between ease-of-use and availability of sophisticated features. It is a long way from Texas to California. in some cases. except. the type of work they are likely to do in future (applied/academic. see Weitzman and Miles 1995). and training. because staff would have to be found to evaluate and then support the new software. Originally users found product information elusive and distribution amateur. Graduate students were particularly likely to be referred back to their own department but regardless of status the general response was that no one had requested such software before and resources were not available. Coding and analysis are (and should be) intrinsically time- consuming. Academics who sought advice from their university’s computing centre almost invariably reported that staff had never heard of CAQDAS. the apparently banal matter of availability. Yet packages vary and users need to choose carefully. Users expected three main things from CAQDAS: that it would save time. and support analytic procedures that were impractical manually. for example. I noted that choice is often ad hoc. policy-related/conceptualizing. Would-be users who finally got the software they wanted did not usually do so by comparative evaluation and discriminating choice. Thus. Applied research features tight deadlines and has relatively straightforward analytic requirements. While applied researchers faced fewer problems justifying acquisition. However. AUTOMATING THE INEFFABLE 169 Take. ignorance was compounded by hostility to the idea of ‘mechanizing’ the craft of qualitative analysis. taking account of not only cost and the availability of sup- port but also the nature of the analytic project(s) for which the software is needed. computing centre resources. The same story generally prevailed when requesters approached their own departments for help. make analysis more thorough. Users complained that data entry and setting-up occupied a disproportionate time relative to the analysis the sponsor wanted.

generally. Computers do not forget or mislay things. INTERPRETATION AND ANALYSIS quicker. Academic users were. Graduate students were most enthusiastic about the conceptualizing features. and. when we invited users to connect the ele- ments of grounded theory procedure to the support offered by particular packages. but reported problems learning how to use them. and conceptualization emergent from the data rather than from formal theory. sometimes they simply did not have time to code all the data. though some appeared to ease their frustration by promising themselves they would use the more advanced analytic features in their own time at some future point. most often ‘grounded theory’. which are . It goes little further than that CAQDAS packages are based on coding data segments and writing analytic memos. enthusiastic about the conceptualizing features. the elevation of ‘closeness to the data’. The value of CAQDAS to them was as an elec- tronic filing cabinet. the software can often help retrace steps. Users tended initially to declare affiliation to a particular analytic approach.170 GENERALIZATION. as these are asso- ciated with more formal approaches to qualitative data analysis. However. computing issues. citing the appeal of rigorous analysis without quantification. accounts were seldom based on direct adherence to grounded theory procedures. Similarly. if users need reminding of how they have carried out some operation. What about the expectation that analysis would be more thorough? Users gave abundant testimony to this. they wanted something more flexible than that. To suggest that CAQDAS particularly supports grounded theory work is as plausible as suggesting that all qualitative research is grounded theory. they always do the same proce- dure the same way. Recent versions of qualitative software offer features which enable kinds of work which were not previously supported. Users were in no doubt that CAQDAS can manage voluminous data more effectively than can paper-based methods. seriously limiting the kind of analytic work possible. Thoughts on the use of CAQDAS in the analytic process prompted reflec- tions beyond technical. in light of the limited use of the more sophisticated analytic features noted in our research. Indeed research can be both rigorous and illuminating without using a ‘standard’ approach at all (Weaver and Atkinson 1994). But evidence was mixed on the support for sophisticated conceptual work. several emphasize the need for thoroughness in the coding process and reflect this in program design (this is not to say procedures for applying codes should not be simple and quick). Thus. Nor is the software as closely informed by grounded theory as the developers’ early claims suggested. It seems that users did not expect CAQDAS to support each element of grounded theory. I already noted that some applied researchers found the pace of their work denied them time for creative play with advanced features. It is particularly intriguing whether these features are being exploited. perhaps suggesting a more open mind and fewer preconceptions. the extent of use of QCA and FCA would be of great interest. asserted affiliation to grounded theory should not determine our view of how qualitative software supports qualitative analysis.

NEW TECHNOLOGIES = NEW ANALYSES? While developers foresee an increasing integration of software tools this is unlikely to result in a ‘superpackage’ having every feature offered by any package. But improvements are continual and prices should be set against the cost of transcription services. Instead. and Kurzweil Voice still pose considerable requirements in train- ing the recognition facility to the user’s voice and checking output for accu- racy. is improving. there will be some standardization of procedures. While technological developments contribute to change in approaches to analysis. Several developers are adding format- ted text features. allowing inferences from tone of voice as well as the words uttered. Adjacent technological developments also have implications for qualita- tive research. conceptual mapping in another. Packages also increasingly provide facilities to export worked projects to websites for joint work and dissemination. where time does not permit transcription. An alternative is ‘direct tran- scription software’. where speech is recorded on a CD-ROM and codes are applied not to text but to the sound segments themselves (e. The new technologies available to qualitative researchers cannot but make for innovation and experiment (and an accompanying degree of confusion). the general move to provide full suites of Boolean operators. modulation and other sound features can be monitored. The latter also supports coding of graphic images. VoiceType. During retrieval the researcher can lis- ten to the actual data as well as read them as text. Code-A-Text. which converts speech to text. allow code assignments to stay linked to text as projects are moved between packages. To enable this. and better-designed interfaces with increased use of ‘drag and drop’ functionality. C-I-SAID. users will increasingly be able to transfer projects between packages. There is also a tendency to more visually based analytic and data presentation strategies. Qualitative Media Analyzer). Such software may also be useful in market research.g. expensive and subject to the sensibilities of transcriptionists. AUTOMATING THE INEFFABLE 171 of considerable interest to researchers well outside the qualitative camp (Lindenberg 1998). or in making a first high-level pass through the data to identify broad themes. a trend encouraged by operation under Windows. with coding done in one package. and so on. but their transfor- mative potential (Lee 1995) is not so strong as to change the external forces . prone to error. nor are they cheap. and accommodate accents and non-standard symbols. N-vivo’s visual approach to modelling. Other changes include more automated coding. Voice recognition software. and the ‘network views’ of Atlas/ti. It is worth noting that the developer of Code-A-Text uses it in analysis of clinical/therapeutic interviews and in con- versation analysis. Current examples include the tree diagrams of NUD*IST4. Software like DragonDictate. to preserve the cues to context these can offer. Data transcription is laborious. Pitch. they are but one influence among many.

The new research technologies are not the source of new analyses. perhaps because such work is more keenly attended-to out- side qualitative research (for example. The network of developers. notably the Internet. The relativist and postmodern positions are products of the Romantic tradition in philosophy (Strong and Dingwall 1997). Both formalism.172 GENERALIZATION. Until recently. These themes are pursued below. INTERPRETATION AND ANALYSIS which shape the qualitative research world or the internal variety which makes qualitative work so interestingly diverse. However. it has prompted more stringent criteria and procedures of quantitative analysis. for example. THE POTENTIAL TO GENERALIZE AND THE INCLINATION TO PARTICULARIZE Critics of qualitative research cite its lack of formality and cumulativeness. and to help those who are so inclined to seek new community standards of partici- pation in the products of qualitative research (see Mann and Stewart 2000). if we narrow our attention from the macro level. in rational choice theory and social simulation). As statistics has developed new tools. such an orientation influenced the work in Event Structure Analysis (Heise 1991). One might cite work by mathematical sociologists in modelling behaviour using ethno- graphic data (Abell 1988). formalizing influences have concentrated on principles for data collection. users and methodologists associated with qualitative software has provided an impetus to formaliza- tion of analytic procedure. These traits compare unfavourably with the formal and systematic character of statistical analysis and survey methods. some methodologists have sought to make the object of qualitative analysis the identification of formal conditions which capture. does allow those who are so inclined to further develop the progress qualitative research shows toward greater formalization. Further. are relevant here. this has been amplified by inattention to analytic principles and procedures in the qualitative methods literature. While analysing words is different from analysing numbers this is not a warrant to be evasive or mystical about our analytic procedures. though they may be their vehicle. the drive to construct algebraic expressions of axiomatic knowledge. qualitative software and other IT. they have not come from technological developments but from adherence to intellec- tual traditions which have grown up in the theory and practice of qualitative research. cause and effect relationships. In qualitative research. Yet. and one impetus to ‘emotionalism’ came from applied studies in medical settings. We might take for example the ‘QCA’ software . and formalization. the effort to codify methodological procedures. The point about these ruptures relevant to qualitative software is that. Technology increasingly facilitates that kind of analysis. while these commitments colour our attitudes to new technology. it attracts less notice than the more vocal analytic orientation which celebrates relativism and repudiates the search for formal knowledge and generalization.

Weaver and Atkinson (1994) pursue an ever-changing. The QCA approach is directed to for- malism and a positivist understanding of data as facts that can be abstracted. In fact. The liberal use of the prefix ‘post’ suggests that these are rebrand- ings of familiar branches of qualitative analysis. at least. 12). but it would not represent a new logic. This is not to say that CAQDAS does nothing new. it transforms procedure. if its proponents are right. The kudos associated with discovery of the ‘new’ encourages authors. It allows us to opera- tionalize procedures and approaches to analysis whose logical possibility was identified but whose demands were entirely impractical before the computer. These approaches share no ground. software is depicted as ‘an approach’ in itself (in the original Denzin and Lincoln volume (1994). standalone chapter on ‘software’. Any ‘transformation’ that emerges will come as a result of being able now to test the consequences of procedures like QCA which. The former addresses causal analysis. editors and publishers readily to assign self-labelled new approaches to a new category. Notice. along with post-positivism and the evergreen inter- actionism. for instance. see Bryman 2001). like Denzin and Lincoln’s (2001) handbook. the latter celebrates chance discovery and the mul- tiple meanings any text can support. Why. one wonders. in itself. Yet both have been operationalized by qualitative software in a way that would not otherwise be possible. may enable credible claims to derive causal explanation inductively from non-numerical data. We still confront the obstacles that were apparent to Weber. No wonder those unacquainted with CAQDAS think it is a new and technicist approach. There will be a difference of degree but not of kind. are software implications not discussed in relation to each of the fieldwork topics and analytic procedures? Instead. It is most unlikely that qualitative software will. An example might be the ‘system closure’ concept in NUD*IST4 and N-vivo (for a good practical account of using the latter. that. even a transformation. the volume you are reading has one. we have several postmodern approaches to analysis. when interpret- ing the data. New technologies even mean we . make it apparent that the growth and subdivision of schools of analytic thought has proceeded quite independently of IT developments. self-conscious ‘landmark’ texts. suggestive analysis and a postmodern conception of data as infinitely contested interpretations. p. like the Denzin and Lincoln (2001) hand- book. That would undeniably be a big thing. AUTOMATING THE INEFFABLE 173 developed by Drass (1989) and the hypertext-linking approach Weaver and Atkinson (1994) document in their writings on the use of GUIDE. where technology transforms social science. feminism and the rest. An interview administered by Netcam allows a researcher in Guildford to do ‘fieldwork’ in Melbourne but the researcher will still have to wrestle with rapport. to put alongside critical theory.1. it is a ’method of analysis’: see Table 1. prompt the invention of new methods or schools of thought. So far. or its lack. If we accept Denzin and Lincoln’s classifications. It should therefore be said that there is little at the level of epistemology or approach to analysis in qualitative software which is not derivative of think- ing that existed before a mouse was anything but an animal that squeaks.

174 GENERALIZATION. Like the date with an old beau. Software which enables individual inputs to be traced via ‘audit trails’ and work to be exported to websites for comment by dispersed team members can encourage sharper thinking and more clarity about how and why a particular analytic decision has been made. as these selfsame funding bodies recog- nize the failures and limits of quantitative research. In some cases. and what may eventu- ally be prevalent uses may not be the obvious or the intended (e. the Netcam example suggests that. Where such features are used to encourage participation in analysis by interested outsiders they can enable debate which enhances the analysis and brings the findings to wider audiences. FACILITATING TEAMWORK. such as that of FCA. which may help us weigh up their impact. If the research subjects are encouraged to use these means to par- ticipate in the stages of inquiry subsequent to data collection we might even see a change in community understandings of research and what it is for. Software will enable us to tag and recover instances where those effects intrude on our analysis. be they government funding agencies. New enthusiasms are volatile. because it makes it easy to review the effects of sequences of program operations and to recast elements of the analysis to gauge their effects. though. in assessing context effects. With technology we can do more. the tele- phone was originally envisaged as a broadcast medium). The strong recent emphasis on relativism has given matters to do with reflexivity extra prominence. INTERPRETATION AND ANALYSIS have to negotiate new ramifications of those problems. A virtue of team research is that it forces individuals to be more explicit about their reasoning. ENCOURAGING REFLEXIVITY It takes time for markets to understand technologies. Again. The analogy contains a warning. the idea of collabo- rative interrogation of the data is the essence of the software design. public service organizations or the private sector. There is a particular potential for hypertext-based projects here. for instance. we will have to make inferences about the effects of not being physically co-present. we increasingly recognize the need to track the effects of our own subjectivity and bring it to bear in our analysis. Qualitative research may enjoy a . and to retrace our analytic steps using ‘audit trails’. CAQDAS packages offer a number of features to support joint work. This helps establish credibility with research sponsors. One non-obvious use of qualitative software that could be very significant is in facilitating working cooperatively. Qualitative software appears to have contributed to the rehabilitation and wider use of qualitative methods. CAQDAS can play a part.g. where out- siders might use the technology to construct their own interpretations of the data (Weaver and Atkinson 1994). qualitative research has also enjoyed new relationships on the rebound. but we also have more to do. by affording the technical appearance of formality that computer procedures have long lent to statistical research (Agar 1991). ENABLING DEBATE.

For methodologists. but our audience is the research community. informed by a particular view of what research is for. This approach also often sees qualitative research as a means of empowerment of research subjects. and features which invite us to extract the maximum from our data. opens our work to discussion with others. provoked new methodological debates (such as those over the status of coding or grounded theory) and tested new analytic approaches (such as qualitative comparative analysis or using hypertext as a freeform method of discovery). It may be knowledge for-its-own-sake or it may be applied. give voice to those whose account is normally silenced. its real significance does not lie in technology or technique but in exposing areas in which accustomed reasoning is muddy. The old prejudices remain. to a website means that others can play a part. qualitative software augments the claim of . Technology does not obscure sloppy thinking. The transparency of the analytic process. or the techniques are ‘non- transferable’ (cannot be taught). At minimum. qualitative research will be back out in the cold. of which research technology is a part. For researchers its bene- fits include data management capacities unquestionably superior to paper- based systems. or our write-up. it exposes it. But regardless of our conception of research. CAQDAS has stimulated a reappraisal of qualitative methodology. are important features of the qualitative scene. where software allows users to recover and display the steps in their reasoning. For example. Research produces knowledge. This may make inquiry more open-ended. Such possibilities do seem to address the business of ‘empowerment’. For research audiences. AUTOMATING THE INEFFABLE 175 new vogue but it is conditional. and make researchers more accountable. AUTOMATING THE INEFFABLE? Qualitative software has contributed greatly to the research community. of course. the user experiences research documented cases where these features encouraged applied research sponsors to play a part in the analysis. In this context. where we seek new audiences by moving away from conventional forms of research writing (Richardson 1994). facilitation of the orderly and accountable practice of analysis. The ability to export parts of our data. research subjects and other participants are enabled to comment on the data and/or the analysis. including those who use research to make policy decisions which affect the commu- nity. We can contrast this with another view. qualitative software can change our relation to our audiences. These views are. If the practice of qualitative research confirms that it is akin to writing fiction or dogma. betrays an inability to decide what the data mean. both the appearance and the reality of formalization. and enabling experiment with approaches which looked promising but whose resource requirements exceeded the capacities of manual methods. It has prompted us to reassess accepted epistemological canons (such as the role of hypothesis-testing in qualitative work).

but. Automating code assignment allows blanket recoding rather than careful inspection of each segment before a code is assigned. It had a place even in survey research as a method to use at the pilot stage. Lincoln (eds) Handbook of Qualitative Research. L. because it was a source of ideas. Neo-quantification of pro- gram output may encourage inappropriate but apparently precise numerical analyses. (1994) ‘Criteria for assessing interpretive validity in qualitative research’. Fielding and R. Those who use qualitative soft- ware testify both that they get ideas from working with the software and that they get ideas in the traditional ways. A. more interestingly. Fielding and R. and Benson. crashes and quirks. (1991) ‘The right brain strikes back’. There are demerits too. and Johnson. Altheide. P. D. in N. in the bath. and even now the chance to make an informed choice and/or to gain access to systematic training is available to few. Lee (eds) Using Computers in Qualitative Research. beckons them to get more involved in the research rather than regarding it as a product they simply consume. London: Sage. in N. Opening up our texts via hypertext-based projects allows untutored use. M. Using software won’t prevent the ‘eureka’ effect but it won’t guarantee it either. in the middle of a conver- sation. (1991) ‘Computing and qualitative analysis: issues in research methods teaching’. Bryman. INTERPRETATION AND ANALYSIS qualitative research to legitimacy. Agar. . and learning from experience (Altheide and Johnson 1994). REFERENCES Abell. Allatt. but by critical peer review. (1952) Content Analysis in Communications Research. Lee (eds) Using Computers in Qualitative Research. Program complexity means that sometimes users have only a vague idea of what particular operations have actually done. which could result in serious disputes with research subjects and the misrepresentation of project findings. Fielding (ed. London: Sage. Denzin and Y. New York: Free Press. liberally interpreted community standards of analytic adequacy. in N. 1st edn. It was ‘a method of discovery’. These and other demerits may be regarded as elements of the one great problem. (1988) ‘The “structuration” of action: inference and comparative narra- tives’. Oxford: Oxford University Press. analyses and interpretations more than they should. For example. J. (2001) Social Research Methods. hunches and hypotheses.176 GENERALIZATION. Qualitative research survived its years as the poor cousin of quantitative social science by creativity. P. Berelson. London: Sage. early users endured endless bugs. not by embracing articu- lated standardized criteria of validity. So what’s new? The field will respond to such problems as it always has. while landing at night.) Actions and Structure. B. in N. that qualitative researchers will claim for their findings. London: Sage.

R. London: UCL Press. Kelle. Lee. (1996) ‘Computer-aided qualitative data analysis: an overview’. Blossfeld and G. . Hoffmeyer-Zlotnik (eds) Text Analysis and Computers. Kelle. qualitative analysis and model building’. (1997) ‘Theory building in qualitative research and computer programs for the management of textual data’. U. (1998) ‘The influence of simplification on explananda: phenomenon- centred versus choice-centred theories in the social sciences’.) Studies in Qualitative Methodology. New Brunswick. F. D. Fielding and R. NC: National Collegiate Software Clearinghouse.) (1995) Information Technology for the Social Scientist. Lee (eds) Using Computers in Qualitative Research. Sociological Methods and Research 23: 86–113. (1995) ‘Confronting CAQDAS: choice and contin- gency’. Lee. N. Lincoln (eds) Handbook of Qualitative Research. Ragin. (1998) ‘Qualitative text analysis supported by con- ceptual data system’. Sociological Research Online 2(2). London: Sage. N. Drass. in G. in R. Thousand Oaks. Lindenberg. N. Raleigh. Heise. Mackensen. in C. Prein (eds) Rational Choice Theory and Large-scale Data Analysis. and Lincoln. Harkness and J. and Dingwall. Blank et al. Current Sociology 44(3): 242–258. Mannheim: ZUMA. London: Sage. (2000) Internet Communication and Qualitative Research: A Handbook for Researching Online. (1987) The Comparative Method: Beyond Qualitative and Quantitative Strategies. U. C. C. Y. London: Sage. unpublished working paper. Hicks. D. 1st edn. and Stewart. Burgess (ed. Miller and R. (eds) (2001) Handbook of Qualitative Research. prob- lems and potential’. London: UCL Press. Richardson. (1995) ‘Computer assistance. London: Sage. S. Dingwall (eds) Context and Method in Qualitative Research. NJ: Transaction. N. Mannheim: ZUMA. in H-P. R. A.) Information Technology for the Social Scientist. CA: Sage. AUTOMATING THE INEFFABLE 177 Denzin. N. New York: JAI Press. (ed. Boulder. London: Sage. in N. (1994) ‘Qualitative comparative analysis and analytic induction: the case for the emergence of the social security state’. Mangabeira. 1st edn. Zull. Strong. in G. Denzin and Y. R. and Lewis. in R. (1996) ‘Diffusion of a methodological innovation: computer-assisted qualitative data analysis in the UK’. E. in N. and Lee. (1994) ‘Writing: a method of inquiry’. R. (1988) Introduction to ETHNO: Version 2. (1998) Computer Analysis and Qualitative Research. L. J. (1991) ‘Computing for qualitative research: options. Berkeley. U. K. CA: University of California Press. Fielding and R. Fielding. Lee (eds) Using Computers in Qualita- tive Research. CO: Westview. and Wille. (1989) ‘Text analysis and text analysis software: a comparison of assump- tions’. in N. P. and Lee. Mann. Heise. (1991) ‘Event structure analysis: a qualitative model of quantitative research’. (1997) ‘Romantics and Stoics’. Fielding. R. and Fielding.M. R. W. and Fielding. K. Lee (ed. (eds) New Technology in Sociology: Practical Applications in Research and Work. Lee. London: Sage.

P.178 GENERALIZATION. Weitzman. Thousand Oaks. (1995) Computer Programs for Qualitative Data Analysis. and Miles. E. (1994) Microcomputing and Qualitative Data Analysis. . M. A. and Atkinson. CA: Sage. INTERPRETATION AND ANALYSIS Weaver. Aldershot: Avebury.

Helen Lucey and June Melody This chapter deals with the intersection of two issues which are central for any research in the social sciences – the issue of the place of our own subjec- tivity in the research process and. We want to explore how we might begin to work with the multiple constitution of those discourses through which the subject is produced. the deeper and more delving our . There is a level at which the practice of data collection suggests that we are seeking a truth about our research participants and that further. out of the intersection of these. It attempts to engage with the place of emotions in the construction of research accounts through our understanding of issues concerned with surveillance. How then can we take the issue of subjec- tivity in research seriously? If research and subjectivity are produced through fictions and fantasies. as though that somehow obviated the necessity to engage with how the intersections of competing fictions and fantasies are lived by the subject. that is. Other chapters in this book raise important issues about the problems of reflexivity. how the intersection of fiction and fantasy is lived for both participants and researchers and how. nevertheless. In particular. certain research stories get to be told. This authorial self is not a core self which somehow shores up the possibility of the account. fictions and fantasies in the research process. like those of the research participants. to examine how this works emotionally.8 SUBJECTIVITY AND QUALITATIVE METHOD Valerie Walkerdine. It is often seen as enough for researchers to assert their own subjectivity without also under- standing the production of that subjectivity itself. the issue of emotion and unconscious processes. then the issue of the narratives of the researcher and participants becomes more complex than the telling of differ- ent stories. issues of subjectivity do not disappear. truth. trying to tell a story about themselves as part of the research. we want to draw attention to Beverley Skeggs’ discussion in Chapter 17 of the way in which the reflexive self is formed through the technology of the bour- geois self in which self-narration becomes confession. That self. in order to avoid problems of speak- ing for those Others. However reflexive researchers might be about their complex relation to the Other. The tricky issue of the place of one’s own subjectivity is not new and has been a central issue for feminist research. Nor is it helped very much by a simple reference to Foucault’s notion of the confessional. is created as both fiction (in the Foucauldian sense) and fantasy. important as that is. inside that.

On the contrary. INTERPRETATION AND ANALYSIS questioning. the sharpening and differentiation of models of mind. psychoanalysis aims for the translation of unconscious fantasy into rational understanding. . the restructuring of methodology. the more profound that truth of the subject will be. Such a decline.180 GENERALIZATION. (Elliot and Spezzano 1999: 28. the rediscovery or invention of psychoanalysis as a vibrant theory and practice. However. Third. while traditionally. that the linear model of the subject is challenged by notions of multiplicity and fracturing. but by fantasy. We would add to this one that is at once social. we argue that what is emerging today is a kind of psycho- analysis of psychoanalysis: a running together of modernist and postmodernist psychoanalytical currents. by defences which prevent one position from spilling into another. 1998). which has effects in an approach to intersubjectivity which emphasizes the link between unconscious desire and otherness. desire and affect’ (1999: 28) and intersubjectivity. postmodern approaches internally critique the desire of psychoanalysis to be scientific and relate that critique to issues of epistemology and interpretation within the social and human sciences. POSTMODERN PSYCHOANALYSIS Elliot and Spezzano (1999) specify this in relation to what they take to be the building blocks of a postmodern psychoanalysis: The development of a postmodern orientation to psychoanalysis is intended to draw attention to the decline of traditionalist. original emphasis) They go on to argue that there are three key points. and the rethinking of interactional configura- tions in which the self is understood in relation to others. subjectivity is not simply about being the ‘sum total of positions in discourse since birth’ (Henriques et al. First. but rather a complex understanding both of discur- sive constitution and the ways in which the relations between positionings are held together by and for the subject in ways which can be quite contra- dictory and conflictual. however. aiming for a model of interpretation which is at once historical and personal. Understanding subjectivity therefore demands an under- standing of emotions not because it seeks to uncover an essentialist depth psychology but because the fictions of subject positions are not linked by rational connections. modernist approaches to knowledge and experience. Second. is not coterminous with its disintegra- tion. the postmodern version ‘underscores the centrality of imagi- nation. This idea is seriously disturbed by Foucault’s writings on the confessional and the idea that such an approach could be part of a will to tell the truth about the human subject – a desire to ‘know’ them psychologically rather than an understand- ing of the way in which they are produced as subjects by the very narratives and discourses that position them in the social world. This works both for the researcher in the unsteady task of uncovering discourses and narratives and for the research subject and indeed for the dance between them which produces the stories told within the research.

debated and critiqued (see Walkerdine 1997 . So how can we explore the implications for method of taking account of both the subjectivity of the researcher(s) and the subjectivity of the participants as constituted through fiction and fantasy? RESEARCH AS SURVEILLANCE The background to this chapter is a long trajectory of work which attempted to make sense of the relationship between the work of one of us. that is the working- class child. as an academic who grew up in the British working class. then the researcher is both written into and writes that story. conflicts and contradictions between them experienced both by the subject and as producing the subject. of empirical work. Our argument is that postmodern psychoanalysis potentially provides one way of understanding these issues by reference to concepts of fantasy. she became aware of the ways in which the families she was working with might ‘read’ her as a middle- class researcher. SUBJECTIVITY AND QUALITATIVE METHOD 181 cultural and psychic. in short. It is important therefore to recognize the way in which psychoanalysis is being used here and what psychoanalytic concepts are referred to. wanted them to be able to see the very part which was painfully invisible. one which recognizes the critiques of psychoanalysis. did not want to look on them with a surveillant gaze. The researcher’s own fantasies become singularly important. defences. in fact. anxiety. albeit in a postmodern fashion. This work was subsequently written up as ‘Video replay: families. Hence our attempt to produce a new methodological turn. who she imagined to be more like them and who. and attempts to move away from any simple depth concept of a ‘self’ (Henriques et al. of interpretation in the social sciences and tries to find a way forward. the storying into being of an account. who were both the object of a surveillant research gaze and the object of her own fantasies about their rela- tion to her. When she was conducting some research on young girls at home and at school for a project in the early 1980s. uses a mode of explanation which removes historicity from the account. Valerie. cultural and psychic are researched together and ways found to develop methodologies which respond to the demand for inseparability at the level of explanation. desire. Moreover. The central issue is to understand the way in which historically specific subject positions are held in place and the relations. of course. This is akin to the ‘third space’ which Cohen and Ainley (2000) characterize for youth research. We would argue that such an approach is perfectly compatible with narra- tive and discourse approaches to the understanding of subjectivity and considerably adds to them (Frosh 2001). films and fantasy’ (Walkerdine 1985) and has been much discussed. many would argue that even keeping a notion of the uncon- scious and working with psychoanalysis. and working-class families who were the object of her research. in which the social. 1998). to unconscious processes. while she. indeed well hidden. if we understand the research process itself as the construction of its own fiction. affect. Of course.

But. In particular. Lucey and Melody 2001). sex and pregnancy. How could she know and yet how could she not take account of both her own fantasies and those of the parti- cipants as these were so highly significant if we were attempting to under- stand not just subjectification (the discursive production of subject positions) but subjectivity? Since then the three of us have attempted to develop this line of thinking more systematically. What we are inter- ested in here is the way in which this expressed ambition sits alongside the fol- lowing extract from an interview conducted by June Melody when she (June) was heavily pregnant. INTERPRETATION AND ANALYSIS for a review). our impossible desire to know what is really going on. but has begun another and met a sympathetic lecturer who is trying to help her in her ambition. June and Sharon are discussing the possibility of becoming pregnant as a result of unprotected sex. 18-year-old Sharon. . how on earth could he know that she was not just projecting onto her research participants her own fantasies that in fact had nothing to do with them? Of course. The conversation in the interview turns to relation- ships. what does this mean for the place of one’s own subjectivity in the research process? If the issue of her own class background was such a potent emotional issue for Valerie in conducting the fieldwork. She has not been very successful and in fact dropped out of one course. what place did it have in understanding the data? In the 1985 chapter. we want to dwell on the issue of the way in which we tried to understand the research process as surveillant – what is it that social and psychological research wants to know about the Other and why. The discussion which follows forms an attempt to begin to take seriously what it might mean to use psychoanalysis to understand the subjec- tivity of the researcher as it intersects with the participants and to argue for taking the fantasies and defences of participants seriously. an ambition her working- class family knows about and endorses. This ambition is at one level obviously a laudable post-fem- inist one. he was quite right. we discuss the issues of transference and defences as they relate to attempts to take the place of subjectivity seriously in a systematic way. who did poorly at school. Sharon has begun a new relationship with an older man with whom she is having sex without contraception. Here is one of the participants.182 GENERALIZATION. THE SUBJECTIVITY OF THE PARTICIPANTS We want to begin by making reference to a body of work contained in our study of transition to womanhood in Britain (Walkerdine. to think about both the tricky surveillance of the research endeavour. Sharon has tried to undertake a BTEC (an alternative to A levels in Britain leading to university entrance). but she is a very long way from achieving or even knowing how to achieve university entrance to an undergraduate degree. to get up close and the importance of understanding our own place in research. as one critic clearly argued. associations and identifications with them. but harboured the ambition to be a judge. she speculated about meanings made by the working-class participants on the basis of her own fantasies. In particular.

) J. but so I mean you don’t – if it happened it happened kind of thing. the mother.M. The researcher has also to . it doesn’t cause you any anxiety – Sharon: I – me and my mum said to me if it happens.M. it is possible to understand the probable yet unplanned pregnancy as being produced at the intersection of two compet- ing positions – the post-feminist judge and the working-class young mother.: Right. the second and related issue is that the research participant is one part of the production of this in terms of a research narrative. it happens. If Sharon retains the fantasy of becoming a judge.M. if you’re not using any contraceptive.: You don’t – not at all? Sharon: No J. without ever having to apparently give up on the ambition to be a judge.M.: But if you’re not using contraceptives then it’s very likely that you will.M. J.M. J. a position so much easier to deal with for her family and well known and trodden (as her mother admits). And you what – do you use contraceptives? Sharon: No J. J. This interpretation demands an understanding not only of the contradictions between multiple subject positions but also of unconscious processes as a place where such conflicts and contradictions can be appar- ently worked on.: And are you hoping that you will? Sharon: No not really. J. Sharon: That’s the chance innit.M.: Right – so do you think you’re trying to get pregnant? Sharon: No J. Just cross that bridge when we come to it.: So what – he doesn’t use any contraceptives? Sharon: No Here June is finding it incredibly difficult to come to terms with the idea that at some level Sharon is not trying to get pregnant. Um Sharon: S’pose I’d be scared if I didn’t have my mum and dad’s backing.M. We want to discuss this extract in two ways. Sharon: (untrans. This relates to the issue of how sub- jectivity is lived that we mentioned at the beginning of the chapter. Sharon: Yeah it’s likely that I will but I’m not like going out of my way to get pregnant or nothing like that. But you’re not avoiding it so – um and is your boyfriend quite happy about that possibility? Sharon: Oh he don’t he don’t want me to get pregnant. J.: Right.M.: Yeah sure.M.: Right – so it’s quite likely that you will.M.: Right. First. a fantasy so very hard to live out. SUBJECTIVITY AND QUALITATIVE METHOD 183 Sharon: It’s the chance that you take though in’t it.: On some level. Really J.M. resolved or kept at bay. J. However.: So it’s possible that you might get pregnant? Sharon: Yeah J. Hope in a couple of years I will but not yet. the unplanned pregnancy could be understood as offering a way out to another known and sanctioned position.

It seemed that she would do anything rather than give me a straightforward answer and that she was hoping to give me as little information about herself as she could get away with. Possibly in the hope that I would not be able to build up an accurate picture of her. feelings of anger and irritation. Sharon’s response brought up some very difficult feelings.184 GENERALIZATION. Of course. And indeed. Adding the researcher’s voice in most cases is designed to fill some of the absences which ‘difference’ produces in order to construct a more complete. including ethnographic research. (Cohen 1999: 11. In fact I really don’t think she wanted to be interviewed at all. I wanted to tell her how bloody stupid she was. original emphasis) . which she expresses in her field notes: I felt incredibly irritated with her. Most qualitative research. nor is the real simply a site for a projection of fantasy. anxious and ‘defended’ (Hollway and Jefferson 1997). By the same token the imaginary is not a distorted reflection of the real. In many ways. As Cohen (1999) argues: the relation between the real and the imaginary is not fixed. when attempting to take account of unconscious processes which are set in motion by all kinds of anxieties and fantasies. more ‘real’ ethnographic picture or which turns to the researcher because of the problems of Other discussed earlier. However. While research in the social sciences is overwhelmingly premised on the notion of a rational. In that sense then. is infused with a realism. the subject of our discourse is alto- gether more irrational. the idea of multiple subjectivity does not simply reside with the participant. INTERPRETATION AND ANALYSIS make sense of this conflict. For June. any notion of what constitutes the ‘real’ is seriously challenged. in many ways the classic aspect of research as surveillance. How might we then make use of this multiplicity of both participant and researcher in the production of the research story? We want to suggest that this takes us beyond a notion of reflexivity and towards this place in which the conflicts and contradictions of a multiple subjectivity are held in place. We need to take account of the fantasies of the researchers and subjects in producing any account. calculating subject. I don’t have any strong feelings about whether she gets pregnant or not. I felt she was being quite hostile towards me (or was I feeling hostile towards her?) and was very ambivalent about being interviewed. the copious discussions of ethno- graphers’ differential treatment of their own field notes and their formal pro- duction of an account makes that ambivalent desire very clear. but tactically deter- mined. with claims to an authenticity which purports to ‘tell life how it really is’. in this case. as interviewer. we can understand this dance as a classic relation of research desire – the desire to reveal counterpointed by the desire to conceal – on both the part of the researcher and the researched. It was nothing to do with morality. June’s desire to ‘build up an accurate picture’ of Sharon can be counterpointed by what June projects onto Sharon as a desire to avoid this. We are always dealing with a process of double inscription whose articulation varies according to a range of social circumstances.

Jennifer Hunt (1989) examines the methodological implications of a psychoanalytic perspective for ethno- graphic fieldwork. However. In the field of urban sociology and cultural geography. Frosh 2001). Rose 1993. What is being . pointing to the issue of transference and counter-transference in fieldwork. In psychology. Detachment is often a form of defence. we suggest that social and cultural analysis desperately needs an understand- ing of emotional processes presented in a way which does not reduce the psychic to the social and cultural and vice versa. Oral historians have combined the techniques of life story research with insights from family therapy in order to explore the ‘mixture of conscious and unconscious models. a number of researchers have begun using psychoanalysis (for example. Pitt 1998). In the volume Psychoanalytic Aspects of Fieldwork. While the overwhelming majority of educational research is concerned with conscious processes. Sayers 1995. Raphael Reed 1995. by examining the issue of projection of the subjects onto the researcher and vice versa. with notably few exceptions (Raphael Reed 1995. However. She pays particular attention to the psychodynamic dimension of the research encounter. there is a growing and significant body of empirical work which is con- cerned to explore individual and institutional patterns of investment and disavowal which enter in the formation of pedagogic identities (Britzman 1995. myths and material inheritance’ (Bertaux and Thompson 1993) which combine to shape individual and family narratives (Ginzburg 1990. SUBJECTIVITY AND QUALITATIVE METHOD 185 The use of psychoanalytic concepts to theorize social phenomena and processes is growing in a number of disciplines. Aitken 1998. Using psychoanalytic techniques and theory in research involves using ideas that have been developed in the context of individual analysis and applied to something that is not taking place in the analytic context. society and space (Pajaczkowska and Young 1992. While we would not dispute the undoubted and well- rehearsed problems with the universalism of psychoanalytic theory. This can be both comforting and distressing to researcher and researched alike. Shaw 1995. Thompson and Samuel 1990). DETACHMENT AND DEFENCE Usually the researcher listens to a story or makes an account but has no place to intervene or do anything with the account except to produce an academic narrative. Pile 1996. Kvale 1999. but recognizes their mutual imbrication. researchers and writers are drawing in particular on the work of the object relations theorists in order to explore the relationship between subjectivity. Hollway and Jefferson 2000) there is little sociological engagement with the intrapsychic dimensions of research methodologies. The essential feature of using psychoanalysis as a research tool is that the researcher is the primary instrument of inquiry. Cohen 1999). creating boundaries is quite different from being a detached observer.

When Valerie did that early research in the 1980s. Within psychoanalytic theory. But. was the object of a painful recognition. Her own attempt to deal with this was to write about her own fantasies that they could recognize her as the working class girl who was just like them and in turn that she could recognize them by being able to identify with them. That which was gone forever. she simply used her own conscious identifications with the material they gave her. Telling her own story therefore became an important way of explaining why she made a par- ticular interpretation. The feeling that I am like my mother can signal a covering over of the painful differences between us. she was trying to sit at the difficult place of an acade- mic who had grown up in the working class researching a working-class family who correctly understood research as surveillant. Of course. identifications and fantasies do not disappear when we are engaged in research. This emerged because they seemed to have so much free- dom of choice and both of us found that we felt particularly angry and envi- ous of the way in which middle-class mothers would ask their 4-year-old daughters what they wanted to eat. When Helen and Valerie researched middle-class families in the 1980s we became aware of our own envy of them and indeed what we saw as our hatred and contempt for them. as Elliot and Spezzano (1999) state. the subjects. research as an activ- ity of Cartesian rationalism is itself to be critiqued. more than this. indeed. or my feeling that I am nothing like her can defend against feel- ings that I am far too like her for comfort. We had not been offered a choice of . but had moved to a different social location and could no longer be recognized as that subject. INTERPRETATION AND ANALYSIS defended against are intrusive feelings about the research process. In this case then.186 GENERALIZATION. and the relationship between the two. the critic correctly asked how indeed she would know that her interpretations were not simply projections of her own fantasies onto the family. We want to argue that transfer- ences. identifi- cation or its absence can signal a defence against its opposite. except in fantasy. But as we said. instead of deal- ing with the complex and ambivalent position of her own fantasies and their relation to those of the participants. it now seems to Valerie that in fact her own desire to identify herself as like them and them as like her was a painful defence against the recognition that she was indeed not like them at all. one irony of this was that she could speak of her own working-class childhood only as an academic who now had the legitimate space in which to speak. including issues of transference and counter-transference (Walkerdine 1997). But what would it mean to be able to separate out the fantasies of the researcher and those of the researched? Do the issues of transference and counter-transference have anything to offer in storying subjects into being? It is all very well to speak of the ‘defended subject’ (Hollway and Jefferson 2000) but what of the defended researcher? The story Valerie told in the 1985 chapter about the working-class family and their watching a video of Rocky II was a romantic narrative in the sense that it tried to make an uneasy case for speaking about her instead of trying to speak for them. It is therefore crucial to acknowledge and attempt to understand what transferences and counter- transferences might be telling us as researchers.

giving Mrs Falmer space to cry. SUBJECTIVITY AND QUALITATIVE METHOD 187 meals as children! That something so apparently trivial can invoke such intense feelings must be a commonplace feature of the research encounter.’ The researcher can then feel some grati- fication about having something to do with this process as well as having her fantasy of being a therapist partially fulfilled. In this sense then. The issue here then is when those positionings are sedimented in the account as fact rather than as shot through with fantasy. The researchers and researched may like or resist their positionings. It could be argued that these posi- tionings have multiple effects. June was very worried that things might become more difficult for Mrs Falmer after the interview and that a ‘can of worms’ had been opened and she was partly responsible for doing this. whether the participants projected them onto the researchers or whether they were mutually created. There is not just a simple story to be told. for example. In her field notes she wrote: ‘I imagined that having a good cry was what she needed. and produce fantasies on both sides. allows them to be mobilized as ways of coping with the encounter and defending against difficult feelings. However. In order to interpret these feelings. Perhaps what is important is that it is these positionings through which the subsequent interview is conducted and understood. relieved that she was accepted as someone who would be ‘safe’ to cry in front of. while June felt positioned by Mrs Falmer. not saying too much. . that it was good for her and was probably healing in some way. we can understand that the positions of. June felt trapped between being an ‘observer’ and not getting involved and wanting to comfort and reassure Mrs Falmer. This is a long way from the presentation of an autobiographical account or of a reflexive account as some- how true because aspects of the researcher’s narrative have been included. for the interview to be therapeutic. Both researcher and researched can then use these fantasies to position each other within discourse. we need to make some connections with the sub- jectivity of the researcher. researcher and researched. a professional middle- class mother. However. The story is created in and through these dynamics and so it seems important to understand what the dynamics are – and not only for the subjects of research. In the case of Mrs Falmer. But she was also worried that the interview might be abandoned and was relieved that Mrs Falmer wanted to carry on. In the transition to womanhood research Helen felt positioned by working-class Mr and Mrs Cole initially as if she were a welfare agency ‘snooper’. neither does this suggest that the subjectivity of the researcher is unimportant – far from it. as if she were a therapist. It is difficult to tell whether the researchers projected these positionings onto the participants. who cried during the interview. June felt at first comforted by the therapist positioning and then suspicious of her own desire to inhabit that position. not intrusive. ‘benefits officer’/’welfare snooper’ or ‘therapist’ are both created in discourse and work because the fantasy life of the subjects. At this time June was taking steps to begin a psychotherapy training and fairly spontaneously took up the role of thera- pist.

transference and counter-transference are not liter- ally translatable to the interaction with the researcher and subject. usually from the patient’s childhood. interviewers were occasionally left feeling depressed after an interview even when the interviewees presented themselves as happy. Her mother has carried extreme feelings of anxiety about Heather’s intel- lect since this time. However. that the patient imposes on the analyst. the difference between the sisters is exaggerated into a huge gulf. Heather. INTERPRETATION AND ANALYSIS TRANSFERENCE AND COUNTER-TRANSFERENCE Because of the complex nature of fieldwork and the fact that it does not take place in a clinical setting. ‘not very acade- mic’ and a ‘slow learner’. They variously describe her as ‘not very bright’. She went on to Oxford to study medicine. Heather’s exam results nevertheless placed her within the top 10 per cent of the 18-year-old national cohort at that time. within the family. BEYOND CONSCIOUS IDENTIFICATION The following example of one family with two sisters presents for us a way of understanding a possible way beyond the issue of conscious identifica- tions with the participants by the researcher which we mentioned earlier. Angela is seen as the success of the family while Heather. or to the patient’s trans- ferences. She will not be going to university. However. Heather is also a gifted musician and has won much praise for her musicianship. . Counter-transference can refer to an analyst’s unconscious response to the patient or to significant people in the patient’s life. when she was convinced that Heather had suffered brain damage through oxygen deprivation. though there was no medical evidence for this. In the clinical set- ting transference is generally defined as unconscious archaic images. who ‘only’ obtained eight GCSEs and two A levels (all at grades A to C). as sometimes indicating the presence of emotions which have to be projected outwards by the patient onto the ana- lyst because they are too painful or difficult to be experienced by the patient. Although not as outstanding as Angela’s. is viewed as a ‘failure’ by her sister and both parents. and three As and one B at A level. despite the reality of her achievements. Angela has a younger sister. for instance. the research team found the concepts immensely useful. when experienced as those of the analyst rather than the patient. Psychoanalysts understand these emotions. Angela is a middle-class 21-year-old high achiever who went to a presti- gious private girls’ school where she obtained ten grade A GCSEs. In the transition to womanhood research. cheerful and always positive. We learn from her mother that Heather’s identity as not very bright had germinated from the moment of her difficult birth. even though her grades would allow her to.188 GENERALIZATION. We may feel a strong emotion when we hear a story but the participant may not seem to react in this way at all. or vice versa. but to a catering college to do a diploma in hotel management.

what are the absences in the narrative and where are the silences. Regardless of the question whether members of the same family can be regarded as independent sources of data for the purposes of arriving at some ‘truth’. listening and later reading for the overall plot. not only within this particular family but also across all the middle-class families. one containing events. connect and disconnect. We could argue that it is partly this fear that has produced ‘Heather as slow’ as a fiction: Angela’s brightness is intricately linked to Heather’s slowness. SUBJECTIVITY AND QUALITATIVE METHOD 189 However. Walkerdine. Of course. the family narrative was very persuasive and June. with the impression that Heather. but how is it pro- duced as true. Our starting point is an apparently unproblematic description by some fam- ily members of Angela as bright and Heather as slow. It was only when we analysed the data that the complexities in the family narrative were noticed. whom she did not meet. We pay close attention to words. The fiction. are positioned as needing success to ensure the continuity of middle-class life. it is crucial to acknowledge that in the circles in which only the top per- formance is considered good enough. The issue is not whether ‘Heather was slow’ is true. was indeed an academic failure and possibly had learning difficulties. introjections and transferences that are in play. Heather has indeed failed (cf. images and metaphors. The desire and need for success within this framework defends against its opposite – the fear of failure. what does it mean. But to take this descrip- tion at face value would be to miss the complex dynamics that are going on. Using this method in the case of Heather and Angela. thus becomes self-fulfilling. characters and numerous subplots. where do different parts of the narrative begin. like other middle-class families. information on Heather’s educa- tional failure would have been amply ‘triangulated’ by the testimonies of three members of her family. Common to much qualitative analysis (Strauss and Corbin 1990. Angela and Heather’s case provides a stark example of how particular subject positions can be formed in relation to other family members. Silverman 1993) we can look at the face value of the individual narratives. created out of the uncon- scious fears and fantasies of the family. fade out. we can problematize the family narrative through an exploration of the unconscious projections. Where in this narrative are there inconsistencies and con- tradictions. came away from the interview. what tone and register of voice is used and how does this change? These observations may prompt us to ask. stop. their occurrence and recurrence. who interviewed the family members. ‘Who is it that is speaking and who is being spoken to?’ Although initially persuaded by . at that point without the knowledge of her academic achievements. Lucey and Melody 2001). and to whom. However. ‘Triangulation’ is a method developed within mainstream qualitative sociological research for dealing with problems of validity (Denzin 1978) and refers to the injunction to check pieces of information against at least one other independent source before regarding them as ‘credible’ (Lincoln and Guba 1985). the ‘story’ that is being told. we wanted to use the insights of post- structuralism and psychoanalysis to go beyond this mainstream notion of truth. This family.

very much like my father. and who do they represent for me? Parent. was always seen as the clever one in the family. We asked. It was later when she reread her interview notes and listened to the taped interviews that she decided to get a copy of Heather’s exam results. child? It is by being aware of and listening to the different parts of our own fantasies. who do I represent for the subject. For them. what we tune into in our own histories may be a projection onto the research subjects. In this case. say. In our analysis we attempted to understand the relation between the fic- tion of positioning and the fantasy. In relation to the Angela and Heather example. social worker. but the defences and fantasies embedded in that positioning are crucial for under- standing the production of subjectivities. to which part or parts of me is the subject speaking? Which part of me is responding? In other words. In order to do that we interrogated our own fantasies. We can be discursively positioned as. analysts always check their inter- pretations against the progress of the analysis. In this case. Of course. has some effect on the progress of the analysis. Whereas my younger sister was not seen as particularly bright and has spent most of her life feeling inadequate especially in relation to her elder sister. and was always given the message that I could achieve whatever I set out to. I. This is where we need to check out our own story and be aware of its place. sister. it is much more difficult to do this in research.190 GENERALIZATION. June felt very uncomfortable and discussed her feelings with the team. a good interpreta- tion is one which produces some change or shift. June asked what narratives. . where it is often not possible to return nor to simply present our interpretation to the participants. as Angela was described. it was June’s relationship with her younger sister that helped us to explore the research encounter at this level. positions and fantasies were addressed by the family narrative. Although the circumstances of Angela’s family and June’s family were quite different. These clearly exposed the notion that Heather was an academic failure as a fantasy. In clinical situations. This gives us an important step beyond Foucault’s subjects as fictions. For all of the working-class girls in our study to get eight GCSEs and two A levels grades A–C would be a cause for celebration. also the eldest. Of course. social worker. instead of simply identifying or not with the participants. INTERPRETATION AND ANALYSIS the family narrative. In her field notes she wrote: Although Angela and Heather come from an entirely different social class back- ground from me there are very similar dynamics between the siblings. as is proposed by some feminist attempts to counter power differentials by taking all work back to be checked by the participants themselves. there was enough in June’s history for her to tune into some of the fantasies described by the different members of Angela’s family. June had a strong feeling that there was something wrong. that the narrative didn’t quite fit. that allows us to tune in to the different meanings of our subjects. and to the place in us that responds to any given message. teacher.

. is not only a projection within that particular family but also an effect of the relationship between specific social and cultural norms. and cultural norms in the case of the Asian father notwithstanding. but this. Most importantly. it told us something crucial about the anxieties which underpin the middle-class girls’ success but also how this is masked by a hyper-rationalism that will always construct and promote a rational story. WORKING WITH DIFFERENT INTERPRETATIONS As we mentioned earlier in the chapter. Heather’s ‘failure’. It could be argued that they were both trying to establish rapport. who was a highly educated professional. becomes a fantasy. discourses and family processes. that Jacky wasn’t ‘giving’ her much. Another Asian middle-class father. June’s hunch that there was some- thing that didn’t ‘fit’ in this family narrative threw much light on other feel- ings that were coming up for all of the research team in relation to the middle-class girls. talked about how angry she was with Jacky for feeling proud of ‘pathetic achievements’ such as passing her driving test. then. It is not just the interviewee who puts up resistances to difficult feelings. a 21-year-old working- class woman. middle-class fathers attempted to ‘interview’ the female interviewer suggested that they were feeling vulnerable and their defence against this was to put themselves into a position that they perceived as more familiar. Margaret. questioned the young Asian woman inter- viewer about her family and her language. It seemed that in some cases the researcher was perceived as having the more powerful role and the participant attempted to have an effect on the dynamic. our transcriber. So for instance. But what happens if we explore disagreements about interpretation in terms of the unconscious dynamics created within the research encounter? When June interviewed Jacky. As we will see with the examples of Anna and of Mr Cole. Looking at who reveals what to whom involves complicated plays of power. it was the researcher who was trying to defend against difficult feelings. qualitative approaches have often attempted to copy quantitative methods by finding a way to agree on the ‘truth’ of an interpretation. and felt extremely irritated by this young women. the regularity with which professional. This is neither a simply social nor a simply psychic process but has to be understood as working in a complex psychosocial manner. In one instance a middle-class father who was financially and professionally very successful and who had spoken at length about his own religious interests wanted to know about the researcher’s religious views. thereby reducing their vulnerability. A view which postulates that it is possible for researchers and subjects to be equal. Also we found that some of our participants (mostly middle-class men) felt uncomfortable with the ‘power’ dynamics of the research encounter. she felt that the interview was a struggle. SUBJECTIVITY AND QUALITATIVE METHOD 191 We are working with the premise that our experience of the dynamic at this time can tell us something important about this persons’ relationship to the wider social world. speaking to her in Urdu.

During the early part of the interview Helen felt that she had to struggle to get any rap- port with him. These very dif- ferent interpretations of the same data can be traced back to the differing histories of the research team. Thus. Treating him as if he didn’t exist may have been a relief for Mr Cole. She couldn’t wait to finish transcribing her tapes. their lives as she saw them had no sense of purpose or ambition. Indeed. This itself can be understood as a defence against the fear that they are too alike and that importantly. Helen’s consistent impression of the whole interview. This example also shows the importance of acknowledging and examining our feelings. which were clearly connected to the unconscious fantasies of the researchers.192 GENERALIZATION. at the time and throughout subsequent case discussions was that Mr Cole . Mr Cole’s feelings of betrayal went deep. that something intruded into the family dynamic. When we analysed the responses of the research team it became clear that when Margaret was 21. unlike most other approaches we did not prioritize the pursuit of agreement among ourselves. He said: ‘You can trust a thief. Mr Cole was so angry with his eldest son (who was. INTERPRETATION AND ANALYSIS finding a boyfriend and getting a job. because we were specifically interested in the interpretative processes through which a particular reading is made. Of course. upward mobility has not completely erased the marks of working- class subjectivity. her life was not so different from Jacky’s. Deeply rooted conflicts about the researcher’s sense of being were mobilized in this research encounter. but not a liar. in fact his stepson) that when he stole jewellery and money from Mrs Cole. This usually involved recognizing aspects of our own subjectivity and the way in which our own histories influenced our interpretations. Consider the following example in which Helen interviewed Mr Cole. as did his reluctance to ‘forgive’ his stepson. ‘thought she was articulate and interesting and more giving than June had perceived’. The transferences that took place during this interview made June aware of how afraid she was of ending up like them. Mr Cole threw him out of the house. as it transpired in the interview. It would have been a nonsense to simply stifle that aim for the sake of agreement. the very things that were both essential and crucial to her own identity. Since her teenage years she has tried as much as she could to ensure that her life was as different from theirs as possible. Sharon’s father. the boy’s existence was living proof for Mr Cole that his wife had not always been with him. June saw Jacky’s lack of vision and ambition and her minor ‘achieve- ments’ as being similar to those of her mother and sister. defending against difficult emotions like jealousy that he would rather not have felt. It needs to be emphasized that there were many disagreements in interpretation between members of the research team and often a lot of discussion about our different responses and interpretations. Helen on the other hand liked Jacky. with whom she has difficult relations. The researcher’s defence mechanisms sometimes meant that she changed the subject or pushed the interviewee into another direction when what she found was too uncomfortable for her to deal with. As far as the family is concerned the boy doesn’t exist’. This example shows us quite clearly the need for the researchers to be able to distinguish what is their material from that of the subject.

she suddenly felt very uncomfortable and reacted by abruptly breaking the rapport. Because of some of the painful things that have happened to her family. We spent a long time as a team exploring why Helen changed the subject. It was difficult for her to keep the Cole and Lucey families separate. CONCLUSION: FANTASIZING SUBJECTIVITY We suggest that far from opening up the simple confessional. However. Mr Cole had asked Helen whether she had any brothers or sisters. at one point in the interview. In this scenario. The fact that she didn’t want to continue the conver- sation suggests that counter-transference took place and this family repre- sented someone else for her. Simultaneously however. and wanted to relieve her anxiety and pain around these issues by changing the subject. an engage- ment with emotions and unconscious processes is absolutely crucial for . psychically. as Helen had literally had the door shut in her face. Analysis of the transcript shows that it was his knowledge of her as coming from a large family that allowed him to project some of his own fantasies onto her and make her a ‘safe’ person to whom to reveal some of his history. it was at this point that Helen suddenly stopped Mr Cole in his tracks and changed the subject! How do we make sense of this interaction? It is important to note that establishing rapport in this particular interview was very important. What seemed to happen is that having slipped into a temporary identification with the Cole family. It was during one of these discussions that it was noticed that. what can be seen as counter-transference is the subse- quent shift from not revealing things about herself and being an anonymous interviewer to temporarily losing her identity as a researcher. It is because it leapt over the boundary for her that it became uncomfortable. Ironically. SUBJECTIVITY AND QUALITATIVE METHOD 193 resisted going into any detail with her about what had happened with this son. all Helen wanted to do was get out of the door that was previously shut in her face. while this particular interaction created more empathy between the interviewer and interviewee. events within her own family which resonated strongly with what Mr Cole was now telling her. In an attempt to move away from these difficult feelings and to hold onto her researcher role she changed the subject. she felt uncomfortable when invited to discuss them by ‘outsiders’. and when she was finally let in the house by Mrs Cole she was left standing in the hallway while they sat in the living room and ignored her. She felt extremely uncomfortable and was very pleased that he said anything at all at the beginning. because it was clear that when she replied that she was one of eight children (he was one of seven) his attitude changed noticeably and he became much more forthcoming with information about his relationship with his stepson. She wanted to answer his questions and wanted him to be able to identify with and feel safe with her. meant that she actually didn’t want to hear any more of this painful story and changed the subject. This question proved to be a crucial moment in the interview.

Rather. such as co-authoring with the participants. And yet. makes us look to our work as fictions not truths. we understand this work as dealing with the impossibility of detachment through methodological guar- antees and have tried to find some way to take seriously a subjectivity that always intrudes. it is part of a fundamental critique of methodology in the social sciences. we are con- fronted with the inevitability of the place of power within the account and the way in which our own inscription as researchers produces a deeply uncomfortable gulf between ourselves and our participants. This approach takes us away from an agreed version of events and towards an account which is always necessarily ruptured. This is not. writing a postmodern reflexive account which claims that the research is more about the self than the subjects of the research. It is for this reason that we felt and still feel that we need an approach which can exam- ine how those contradictions and conflicts work to produce subjectivity in a way which takes us beyond the notion of simple narratives or discourses without ditching the lessons learnt from those approaches. This work is part of a struggle to do research differently. Indeed. using . demands to know what we really want to know and what we want to know it for. no matter what one’s best intentions. doing just that funda- mentally critiques psychoanalysis and takes us into the terrain that Elliot and Spezzano (1999) begin to map in relation to a postmodern turn which takes fantasies as a basic building block. It both challenges the possibility of social inquiry. within the research process. The approach potentially is just as deconstructive of the fantasies of the researcher and which bolster our own narratives of our subjectivity. bias. there are things that we cannot remain silent about. Without this. but also the tricky place of emotions. cannot ignore. ours and the participants’. postmodern and post-structuralist critiques of truth make this at best a shaky exercise because it does not make sense to argue that depth methods simply get us closer to the ‘truth’. at the same time. we would argue that discursive and post-structuralist approaches have no way of understanding how subjects live the contradictions of positioning. such a method needs to go beyond surveillance by interrogating the unconscious phantasies and conscious fantasies of the researcher and their place in the research narrative. the demands of imposed fictions or the exigencies of everyday life.194 GENERALIZATION. nor that psycho- analytic interpretations work on their own without their incorporation into a wider social and cultural framework. just as Beverley Skeggs (Chapter 17 in this volume) points to the reflexive self as being formed through the technology of the bourgeois self. Moreover. Nevertheless. However. INTERPRETATION AND ANALYSIS understanding not only how multiple subjectivities are held together. as we have argued. but would never be part of a model which sees research as unnecessary. On one level it would be easy to say that we should simply incorporate the insights of psychoanalysis into qualitative methods. This is not simply about the production of narratives even if some of those narratives are reflexively about ourselves. This gulf produces an anxiety which is hard for the researcher to live with and hence the many ways of attempting to reduce that anxiety. partial. and can never be. frac- tured. Instead.

Qualitative Enquiry 3(1): 53–70. Pile. London: Routledge. Hollway.S.. (1995) Mapping the Subject: Geographies of Cultural Transfor- mation. Y. N. (1999) Psychoanalysis at its Limits. Myths and Memories. (1998) ‘Qualifying resistance: some comments on methodological dilem- mas’. London: University of East London. London: Sage. and Jefferson. S. Hunt. A. C.C. REFERENCES Aitken. Oxford: Oxford University Press. and Thrift. Hollway. L. (1990) Myths. W. T. Kvale. Pajaczkowska. . Henriques. 3. (1997) ‘Eliciting narrative through the in-depth inter- view’. (2000) ‘In the country of the blind? Youth studies and cul- tural studies in Britain’. (1985) Naturalistic Inquiry. Elliot. Qualitative Inquiry 5(1): 87–113.. Finding the Way Home Working Papers no. and Walkerdine. (1996) The Body in the City: Psychoanalysis Space and Subjectivity. J. T. (1999) Strange Encounters: Adolescent Geographies of Risk and the Urban Uncanny. and Young. (eds) (1993) Between Generations: Family Models. P. Journal of Youth Studies 3(1): 79–95. Venn. C. (2000) Doing Qualitative Research Differently: Free Associations. P. Frosh. (1998) Family Fantasies and Community Space. Bertaux. Pile. But that does not mean that we should give up research because some things are important politically to be said even if their telling is so full of contradictions.K. and Ainley. social distance and the place of ‘truth’ within the effectivity of the social sciences. and Jefferson. Cohen. Clues. Lincoln. representation. N. Narrative and the Interview Method. and Thompson. New York: McGraw-Hill. psychoanalysis’. New Brunswick. Hollway. (1999) ‘The psychoanalytic interview as qualitative research’. The issue then is at least twofold. NJ: Rutgers University Press. and Guba. Emblems. V. D. and Spezzano. The anxiety and the practices which defend against it can be understood not as telling us anything fundamental about human subjec- tivity or psychological processes but about power. (1998) Changing the Subject: Psychology. London: Hutchinson. S. Centre for New Ethnicities Research. Our surveillant and authoritative position makes us correctly anxious. Cohen. Qualitative Studies in Education 8(3): 229–238. P. S. Denzin. Donald and A. London: Routledge. W. in J. International Journal of Critical Psychology 1: 28–46. Qualitative Studies in Education 11(4): 535–553. SUBJECTIVITY AND QUALITATIVE METHOD 195 standpoint approaches to side with the subjects and so forth. C. D.. CA: Sage. P. Rattansi (eds) ‘Race’. Urwin. London: Sage. C. C. (1989) Psychoanalytic Aspects of Fieldwork. E.G. London: Free Association Books. (1992) ‘Racism. Ginzburg. W. (1995) ‘The question of belief: writing post-structural ethnography’. 2nd edn. Social Regulation and Subjectivity. (1978) Sociological Methods. J. Culture and Difference. A. London: Sage. S. London: Routledge. Thousand Oaks. Britzman. Pitt. (2001) ‘Things that can’t be said: psychoanalysis and the limits of lan- guage’. S.

Lucey.196 GENERALIZATION. . Culture and Social Justice in Education. J. Walkerdine. (1995) The Man Who Never Was: Freudian Tales. Kaplan and V. A. London: Palgrave and New York: New York University Press. (1995) ‘Reconceptualising equal opportunities in the 1990’s: a study of radical teacher culture in transition’. London: Sage. Strauss. J. (1993) Feminism and Geography: The Limits of Geographical Knowledge. J. V. Griffiths and B. Gender and Anxiety. H. Text and Interaction. London: Virago. V. Cambridge: Polity.. J. Thompson. L. (1993) Interpreting Qualitative Data: Methods for Analysing Talk. in M. Rose. (1989) Democracy in the Kitchen: Regulating Mothers and Socialising Daughters. Troyna (eds) Anti-racism.L. Walkerdine. and Corbin. (1995) Education. London: Chatto and Windus. and Melody. R. (1997) Daddy’s Girl: Young Girls and Popular Culture. Walkerdine. in J. V. D. (2001) Growing Up Girl: Psychosocial Explorations of Gender and Class. and Lucey. (eds) (1990) Myths We Live By. Shaw. C. London: Routledge. Sayers. London: Routledge. Silverman. P. films and fantasy’. Walkerdine. (1990) Basics of Qualitative Research: Grounded Theory Procedures and Techniques. V. London: Sage. INTERPRETATION AND ANALYSIS Raphael Reed. G. and Samuel. Stoke-on-Trent: Trentham. (1985) ‘Video replay: families. London: Taylor and Francis. Burgin (eds) Formations of Fantasy. Donald. H. London: Macmillan.



In basic ways. but it also poses choices about how to study the social world. from formulating a problem to drawing conclusions. this is so.9 OBSERVATION AND INTERVIEWING: Options and choices in qualitative research Kathleen Gerson and Ruth Horowitz Contemporary debates about the practice of social research have focused on distinctions between qualitative and quantitative approaches. is more likely to proceed inductively. These commitments form the common ground on which qualitative approaches are built. Since these choices arise at each stage of the research process. we will consider how qualitative researchers may resolve them in different ways. interviewing and observational techniques offer contrasting ways to approach the social world. Participant-observation. Similarly. requires a more deductive approach to research design and theory building. The contrasts we consider involve fundamental differences in the way that qualitative research is conceived and conducted. Qualitative approaches typically include attention to dynamic processes rather than (or in addition to) static categories. At each stage in the research process. Qualitative research always involves some kind of direct encounter with ‘the world’. and they aim to discover or develop new concepts rather than imposing preconceived categories on the people and events they observe. which Ruth Horowitz relies on for ethnographic fieldwork. but also with the ways that people con- struct. while ongoing interactions within and among groups . often imply- ing that qualitative approaches share a unified set of assumptions and methodological principles. In-depth interviewing. the unfolding of a life history cannot be directly ‘observed’. the method Kathleen Gerson uses to investigate how personal biographies inter- act with social institutions and structures. This variety adds to the power and utility of qualitative approaches. the most prominent of which are participant-observation and in-depth inter- viewing. Qualitative researchers are also routinely concerned not only with objec- tively measurable ‘facts’ or ‘events’. interpret and give meaning to these experiences. whether it takes the form of ongoing daily life or interactions with a selected group. Yet qualitative methodologies offer more than a set of shared epistemo- logical assumptions. They actually encompass several distinct approaches.

Despite the interactive and interwoven nature of quali- tative strategies. our discussion uses a stage model as a heuristic device to isolate the distinct analytic challenges posed by different research tasks. and so forth. a significant advantage of the qualitative approach is its flexibility in allowing the researcher to move back and forth in a cyclical way as the discovery of theoretical insights prompts adjustments in the research design. Reliance on an inductive or deductive approach also stems as much from differences in cognitive styles as from the demands of different methodological strate- gies. Indeed. To the contrary. In practice. and even a debate between us as individuals. In the end. It requires not simply choosing a topic. While focusing on the distinct strengths and challenges of each approach. ‘designing a strategy’. Yet neither interviewers nor observers approach their work in a uniform way. The actual research process typically involves facing problems ‘out of order’ and coping simultaneously with a variety of methodological and theo- retical conundrums. Drawing on our experiences in the field as well as our conversations about those experiences. these methods share a core of epistemological assumptions that make them com- plementary and interconnected. such as ‘formulating a problem’. we explore both the similarities and differences in the research issues that confront in-depth interviewers and participant- observers.200 CHOICES IN CONTEXT cannot be ascertained by posing questions to individuals. we are the first to recognize that the relationship between observation and interviewing is intertwined and mutually supportive. it is not meant to present an overly dichotomous or antagonistic picture of the differences between observation and inter- viewing (see Becker and Geer 1957. FORMULATING A PROBLEM Formulating a research problem is the most important and potentially most difficult task any researcher faces. there are nevertheless important differences in how the research process unfolds. but rather which forms the foreground and which the background. or simply by asking how some apparently understood slice of social . but approaching it strategically. Neither of us would wish to proceed without the tools and insights of each. If our presentation takes the form of a dialogue. The problem can arise in several ways – by locating an unresolved theoretical puzzle. a good qualitative study requires some of both approaches. Whatever the reason for choosing to emphasize interviewing or obser- vation (and any thorough qualitative researcher will use a little of both). ‘life in the field’. and the choice to interview individuals or observe ongoing group processes reflects stylistic preferences as well as theoretical and substantive concerns.1 We present these experiences in the form of a dialogue and divide the research experience into a series of stages. processes such as data collection and analysis are rarely distinct or sequential tasks. Kleinman 1994). The choice is not which method to use. by observing an empirical para- dox.

Macro-social trends thus provide the starting point for formulating a research problem. are crafting new work and family strategies in a post-gendered age (Gerson 2001). OBSERVATION AND INTERVIEWING 201 experience or organization really works. can be solved only by examining micro-social processes as they unfold in the lives of individuals. however. and exceptions can always be found. but focuses instead on how men respond to changes in women’s lives as they endeavour to carve a place for themselves in a social order no longer defined by male breadwinning. When a project begins with an interesting or strategically located research site. for example. investigates how the ‘children of the gender revolution’ have experienced growing up in the nontraditional arrange- ments pioneered by their parents and how they. participant-observation is the more obvious methodological choice. often implies in-depth interviewing. Gender relations and hierarchies thus appear natural and inherent when hegemonic family structures marked by women’s homemaking and men’s bread- winning predominate. To unravel the complexities of large-scale social change. interpreted. Questions about the social construction of gender inevitably arise. in contrast. . it is a principle that clearly distinguishes our methodo- logical choices. and ultimately shaped by the responses of strategic social actors. Hard Choices (Gerson 1985). my research always begins with a set of theoretical and historical questions about the nature. asks how. Individual interviews provide the opportunity to examine how large-scale social transformations are experi- enced. No Man’s Land (Gerson 1993) follows a similar logic. when households begin to diversify and women no longer consistently conform to gendered rules and expectations. it is necessary to examine the intricacies of individual lives. Periods of rapid historical change provide ideal laboratories for uncovering the social bases of relationships and institutions that may remain hidden or seem given during relatively stable historical periods. While these are not rigid rules. however. the structural forms and cultural processes that either sup- port or undermine the social organization of gender become more visible. causes and conse- quences of some important but poorly understood large-scale social or demographic transformation. in turn. is not acci- dental. My current project. Starting with an empirical or theoretical puzzle.2 Interviewing Although in-depth interviewing allows researchers to formulate their research problems in a variety of ways. not surprisingly. At these points in history. both social and individual. The empirical puzzles they raise. The disparate ways in which research problems arise have implications for how they can be studied. why and in what ways women negotiate commitments to family and work amid the new contingencies of post-industrial society. The focus on processes of change.

202 CHOICES IN CONTEXT Observation Participant-observation studies can begin by choosing a site that helps explore gaps in theory (Burawoy et al. whose members tend to be left out of active participation in ‘mainstream’ social and democratic processes. and even members of the same family were pursu- ing different paths. recent sojourns in Spain and Mexico left me well prepared to communicate in Spanish. it became clear that young people held many different orien- tations and expressed them in a variety of ways. I began to focus on specifying the sociological issues only after some time in the field. While my choice to study a Chicano community was in part driven by the research question. Since poverty and low levels of education were essential for my purposes. After several months. While this community fitted in with my general interests. the first need is to find a group or place to study and to pose a theoretically interesting problem. what limits and what helps to create the social worlds of the people? My first project. which then prompted me to formulate the next stage of the project.5 This insight produced a new focus. Both the culture of poverty and structural strain models failed to account for young people’s creativity or for the struggles they mounted and the choices that they made in the face of great obstacles. moreover. Each of my projects has developed out of a general interest in groups located on the margins of society. I chose the fifth poorest area in the city. but also because of its availability and my personal interests. Others dreamed of attending college and were trying to do what seemed neces- sary to attain this goal. Honor and the American Dream (Horowitz 1983) required locating a community in which I could study young people who were mar- ginal to the economic and cultural mainstream. The young men and women. Some were heavily involved in street life. Midwestern American city. espe- cially in the context of a large.3 The preponderance of Mexican immigrants in this neighbourhood allowed me to study a group about which little was known at the time. the same young people acted differently in different settings. If interviewing emerges from a theoretical or empirical puzzle. participant-observation often relies more on the continuous development of questions during the research process. 1991) or choosing a site that offers the chance to observe groups or organizations of specific substantive interest. Finally. and some of those were members of street gangs.4 This choice of a study site thus emerged not just because of its theoretically strategic location. These findings made little sense from the perspective of the two prevailing models claiming to explain the lives of the poor. Whatever the impetus. had different interactional patterns. Finally. it was also based on my fascination with Hispanic cultures. Examples of questions I ask upon entering the field are: ‘What is really going on’ in such groups or communities? How do people make sense of their social worlds? How do they strike a balance between group membership and wider social participation? And. finally. .

also began with a concern for issues of marginality. who had been asked to evaluate several programmes using survey instruments. The process of formulating a problem thus differs for participant-observation and interview studies. Sooner or later. including its convenience as well as its substantive interest or theoretical rele- vance. focused set of sociological questions. While my initial question. all qualitative strategies tend to expose the weak- nesses of homogenizing or overdetermined frameworks and to replace them with theoretical approaches that focus on the social contexts that enable or . the challenge is to pose questions that illuminate larger theoretical and social concerns and to craft research problems that are theoretically central and empirically focused. Of course. thought it would be interesting to observe one programme on a regular basis. Despite the poten- tially large scope of this project. Conclusion These contrasts make it clear that research problems can arise in diverse ways. Whether the method is interviewing or observation. government- funded welfare programmes were being blamed for contributing to the creation and dependency of unmarried teenage mothers. was vague. Moreover. the choice of a site is critical for it will set the boundaries within which questions and theoretical issues can emerge. Fortuitously. In an observation-based project. A colleague. Whether questions are posed prior to or during the data collection stage. This choice does not obviate the need for developing an interesting. but more specific questions can only be developed as involvement in the field proceeds. they actively participate in shaping the worlds they inhabit. I found a programme that invited me to participate (primarily because one of its designers had a mas- ter’s in sociology and had studied qualitative methods). direct engagement in the social world focuses the sociological eye on the interaction between structure and action – on how people are embedded in larger social and cultural contexts and how. OBSERVATION AND INTERVIEWING 203 My next study. In this case. in turn. Taking advantage of this opportunity. all qualitative research lends itself to viewing socially situated actions and their consequences. ‘What happened in the ‘‘black box’’ of service delivery?’. How did government policies translate from national mandates to local pro- grammes and to the delivery of services within programmes? Why do liberal national policies sometimes further dependency and other times promote independence? These questions emerged as I became involved in the organi- zation and familiar with the challenges it faced. a general question may orient the choice of field site. I wished to focus on the face-to-face inter- actions among these teen mothers and their programme supervisors. Teen Mothers: Citizens or Dependents?(Horowitz 1995). several work-readiness programmes for teen mothers were being launched just as I began to formulate this project. it became both broader and more specific as the research progressed. Sites are usually chosen for a variety of reasons.

effective interviews need to guide respondents through a maze of life experiences in an orderly fashion and within a limited period of time. of course. in contrast. interview studies must thus follow a more deductive logic. the goal is to select a group of respondents who are strategically located to shed light on the larger forces and processes under investigation. It might appear easier. Long before the first ‘official’ respondent is interviewed. A carefully constructed interview guide is also needed to collect information in a man- ageable form for later analysis. My research. simply to choose people at one’s convenience and engage them in an unstructured conversation. The sampling strategy must provide an efficient way to answer large ques- tions with a comparatively small group of people. while observation studies can proceed more inductively. this strategy creates more problems than it solves. It is thus necessary to decide in advance which slices of potentially infinite ‘reality’ are crucial and which are interesting.204 CHOICES IN CONTEXT constrain action and the individual actions that shape the organization of social life. at least in the short run. Ryder argued that young adults both reflect and enact new social forms because they are old enough to be making relatively . In developing a research design. In addition. a theoretically focused study needs to choose a carefully targeted sample that is well situated to illumine the issues under analysis.6 Participant- observers. can enter the field as soon as a research site is chosen. In-depth interviews should. In choosing a sample. but nevertheless. Interviewing cannot begin until decisions are made about who to interview and what questions to ask. for example. The research questions can then emerge as field experiences uncover the central issues and problems. DESIGNING A RESEARCH STRATEGY The techniques of interviewing and observation imply different types and degrees of preparation before entering the field. draws its inspiration from Norman Ryder’s classic insight that young adult cohorts occupy a strategic social- historical location that allows them to reveal and enact processes of social change (Ryder 1965). however. Ironically. less important for answering the research questions and resolving the relevant theoretical debates. First. an interview study requires substantial forethought and advance planning. In the long run. always leave room to discover the unexpected and uncover the unknown. theoretical analysis and pretesting are needed to guide the selection of a sample and the con- struction of an interview schedule (or open-ended questionnaire). Interviewing For better or worse. this is more easily accomplished by choosing a theoretically focused sample and developing an incisive and probing interview schedule.

in turn. and class position) pose different dilemmas. While this sampling approach may not yield a strictly representative sample.g. and supportive guide though a process that can be confusing and unsettling. where the need to find people within a specified age range makes random sampling inefficient. college alumni lists or union rolls) but also contain variation in the important social factors under study. under- standable. The coping strategies they develop in response to new social opportunities and constraints. the challenge is to choose a sample that can expose how different social locations (such as gender. these theoretical and practical concerns mesh in the form of the life history interview. OBSERVATION AND INTERVIEWING 205 autonomous decisions but young enough to be influenced by basic social shifts and to take advantage of new opportunities wrought by social change. strategies that rely on some form of self-selection may be the only way to find a sample. techniques especially likely to produce self-selection. After focusing on a set of social experiences as they are embodied in cohort membership. When the contours or loca- tion of a particular group cannot be known in advance. therefore. offer unequal resources. but varies on others deemed important in the theoretical literature (e. should be avoided. such as snow-balling or advertising. race. involves the development and pretesting of a theoretically informed and effective interview guide. When possible.7 In most cases. Within a specified group (in my case. race. usually an age group or cohort). Although the sampling procedure need not be strictly random. shape the future direction of change. As important. My research focuses on this age group in order to examine how individuals on the cutting edge of social change contend with and respond to unex- pected contingencies as they face crucial choice points in their lives.g. it provides a research strategy for discovering the ways that social circumstances shape individual experiences and choices. those who have offered to give their time and share the intimate details of their lives also have the right to expect a clear. In the case of a cohort study. By choosing a sample that controls for one consequential aspect of lived experience (e. however. it is nevertheless important to interview people who vary in their social resources and in their responses to change. and create divergent options. First. age or generation). it is more appropriate to use lists that not only can be sifted by age or birth date (e. The life history format draws theoretical attention . class). the researcher needs to know what kind of information to gather. such as class. gender. it is impor- tant to choose a sample carefully and with as little bias as possible. A successful interview study also depends on the prior construction of a theoretically informed and user-friendly interview schedule (or question- naire). race and gender diversity.g. One of the first phases of an interview study. In my research. selecting a sample randomly from a larger group that shares a set of theoretically relevant characteristics is preferable. the aim is to discover how similar social changes are experienced by different social groups. Readers and researchers alike need to be confident that the findings reflect larger trends and not just the idiosyncracies of a narrow or self-selected group.

Chronologically ordered questions thus provide a struc- ture for recounting a coherent narrative and for remembering potentially important. develop new beliefs and ideologies about how to live and rear their children. Finally. such as child-bearing and child-rearing decisions. family. history and social structure. Using this approach not only helps elicit theoretically relevant information but also helps create a comfortable. and the person’s evolving and current interpretations of the experience. I also examine how people cope creatively with structural dilemmas and. When the inter- view moves through a series of experiences. the social context in which an event or experience takes place. events and situations that trigger or prevent action. in the process. called the orienting framework of social science – the intersection of biography. ideological outlook and sense of self. By focusing on the events. it is not a theory in itself. it provides a powerful method for assessing the usefulness of prevailing theories and developing new theoretical perspectives. (In my work. prefer- ences. but easily overlooked events and experiences. Comparing processes of change and stability highlights the ways that social arrangements can either reproduce pre-existing relations or prompt the emergence of new social and behavioural patterns. In this way. from past to present to imag- ined future. a well-constructed in-depth interview goes well beyond the more structured survey to explore a range of theoretically important dimensions. or circumstances that transform a person’s life path. The life history approach offers powerful analytic tools for developing and using theory. the person’s feelings. in The Sociological Imagination (1959). this framework draws the analyst’s attention to processes of change over time. but rather as the unfolding of events. comprehensible structure within which participants can discuss their lives. the person’s behavioural response. Wright Mills. . perceptions and feelings over time. including pre-existing beliefs and outlooks. Rather. the social contexts in which choices are made.206 CHOICES IN CONTEXT to what C. it is important to include probes that distinguish among the various dimensions of lived experience – including the actual event. during and immediately following the experience. people experience their lives not as a set of factors or variables. I investigate how beliefs.) While life course analysis offers a framework for discovering theory. struc- tural. factors. work or relationship histories) and to place these histories in a social and perceptual context. the focus on unfolding action strategies over time and within a social context helps to untangle causes and effects. and the longer-term interpreta- tions that people develop as their lives proceed. Indeed. change in response to structural and cultural opportunities and constraints. By framing questions in terms of baselines and trajectories of change or persistence. perceptions and beliefs before. and cultural bases of choices and actions that might appear natural or predetermined. it becomes possible to uncover the social. people are able to recall the unfolding of a specified set of occurrences (for example. in formulating a depth interview. the social and psychological conse- quences of contextually embedded choices. Sociological theories aside. and choices.

Despite the dearth of community research being conducted at the time. Gang members tested me on several occasions over the three years in the field. emerge in different ways over time. when class and ethnic tensions and antagonisms made it difficult for university researchers to study poor. In this way. which tend to posit unchanging psychological and even biological causes of behaviour. A careful. I was able to overcome the resistance and become an accepted part of the daily landscape. a participant- observation study can proceed in a variety of ways. Constraints on access to particular sites or groups are typical. These theoretical concerns provide the direction for developing an interview schedule. the interview schedule can direct people to focus on the research’s central theo- retical concerns. and the perceived social distance between the participants and the researcher. And while no one had the formal authority to keep me out. investigated and chal- lenged me. a camera was taken from a newspaper photographer and destroyed. a number of politically active community members interviewed. the degree of control exercised by the participants. I was able to develop close. A month before I entered the community. Since my research is concerned with how and why people make their work and family choices and travel different pathways. there were other means to do so. Observation After a research site is chosen and a general question is posed. for they point toward the critical factors (or range of important independent variables) and the critical outcomes (or dependent variables) that the interviews need to explore. which posit that human actions are embedded in social arrangements. the politics of the group. the cultural and historical context. Problems of access can vary depending on the type of setting (formal or informal). My study of Chicago Chicano youth took place during the 1970s. this task must proceed alongside tackling the more practical challenges of gaining access and settling in at the research site. amid the suspicions that I might be a prostitute. must be to gain access to the study site or group. ongoing and continu- ally evolving relationships with many in the community. with social construction and developmental theories. Yet. OBSERVATION AND INTERVIEWING 207 My work is specifically concerned with comparing the power of individualist theories. The first step. however. It measures both the things to be explained and the full range of potential explanations. respectful . minority communities. and the political environment can make access especially problematic. and have consequences for the form social structures take. To be sure. my open-ended questionnaire (or interview schedule) must focus on measuring the shape of work and family trajecto- ries (from early childhood expectations to current situations and outlooks to future plans and aspirations) as well as the childhood experiences and adult opportunities that might influence these diverse outcomes. for example. a narcotics agent and a social worker. While specifying the socio- logical problem must continue as well.

this con- versation took on a deeper meaning in the context of theories about the social disabilities of delinquents. Entry to controlled spaces. For example. Observational techniques involve a continuous process of observing. life in the field consists of a continuously evolving research strategy. such as prisons. the data-gathering strategy may also need to change. Unused to such gallantry. Starting with a general question renders everything potentially interesting and important. in turn. My attention to that one small event thus triggered a focus on how gang members were able to adjust their behaviour from setting to setting and to make choices that reflected the needs of different situations. a gang member took my arm as we crossed an icy street on our way to the corner store. it is necessary to observe and record almost everything in the beginning. I blurted out. are protected by powerful groups and individuals. A fortuitous encounter nine months later. enabled me to dispute the validity of social disability delinquency theories. . Upon entering the field for my first project. As Becker (1998) suggests. but the research strategy was altered and the research ques- tions reformulated. politics. designing and carrying out a project may depend as much on fortuity. quinceañeras. This is a particularly telling example of how the choice of – and access to – a site shapes the kinds of questions that can be posed and the kinds of answers that can be found. In the end.8 I found many additional inci- dents that cast doubt on these theories. provided entry into a similar organization. ‘My mother taught me to take a lady’s arm when crossing the street. ‘Why did you do that?’ He responded. and study- ing organizations of this type may pose considerable obstacles. When powerful people control access to a research site. I was recently denied access to a professional organization despite their interest in my research and my own position as a group member. Since any observation can ultimately become sociologically relevant. I attuned myself to all activities. it would not have been possible to develop and test sociological categories and their links. Once in the field. as well as small. Without constant observation of the smallest interactional details. and restaurants. The research could continue. for example. developing categories or classes of phenomena and their links. seemingly unimportant actions or words that can become critical. and searching for new data to explore the new categories.’ Although a seemingly innocuous exchange. one enters the field with concepts and ideas that direct our gaze and need to be recorded as fully as possible. powerful members rejected my entreaties to observe closed meetings. In this way. businesses and schools. the strategy was to locate a variety of groups and follow them as they traversed different settings. the obstacles may be more daunting. analysing. This insight. however. networks and personal assets as on theoretical and substantive interests. As I began to follow gang members to weddings. After about five months. Worried about legal questions and vulnerabilities.208 CHOICES IN CONTEXT and determined approach can overcome the obstacles set by initially mistrustful people.

OBSERVATION AND INTERVIEWING 209 Conclusion Interviewing and observation typically involve (and often require) divergent research strategies. Yet the prospect of entering unknown situ- ations. Repeating this process dozens and even hundreds of times requires energy. And whether the method is interviewing or observation. The exhilaration stems from putting a plan into practice and doing sociology rather than just consuming the ideas and findings of others. the most anxiety- provoking task of the interviewer. but some differences in analytic technique are bound to emerge from these differences in data-gathering strategies. the in-depth interviewer must spend more time and forethought preparing for field encounters. fieldwork becomes as much an art as a science. It takes a strong belief in the value of one’s project and a certain amount of chutzpah to ask others to share their most personal. confidence and unassailable commitment. Even though most agree to help. and resis- tance. Securing the help of strangers is. it poses emotional as well as analytic challenges. Not all interview and observation studies need to diverge so starkly. Over and over again. in some respects. the interviewer must contact and secure the help of others. Whatever the degree of advance preparation. The key is to choose the method that fits the problem. engage in an intensive process whose goal is to reveal the intricacies of other lives. the possibility of rejection arises anew each time a new . Interviewing An interview study involves a series of discrete but demanding forays into the lives of strangers. it is both exhilarating and frightening to begin the process of directly engag- ing ordinary people as they go about their daily lives or reflect upon where they have been and where they are going. fear. Interviewing depends on developing a research design from more general empirical and theoretical concerns and thus relies on a more deductive approach. intimate stories for no other reasons than the advancement of knowledge and the possibility of increased personal awareness. persistence. and then say ‘good- bye’. Both nevertheless aim to expand or transform common understandings of social processes and prob- lems. courage. and putting one’s most cherished ideas and perspectives to the test also tends to provoke anxiety. meeting countless strangers. At this stage. travel to unknown places. An observation study can begin as soon as a research site is chosen and then proceed more inductively as what is really going on becomes clearer and a set of more specific questions and insights begin to emerge. LIFE IN THE FIELD: ENCOUNTERING THE SOCIAL WORLD While life in the field can begin almost immediately for the participant- observer.

The structure of this situation. if .10 When these ideals are achieved. support and intense concentration. most are grateful for the attention of a supportive listener and energized by the insights they have gained. rapport and mutual commitment within a short time period. The best interviews become a con- versation between two engaged people. the intensive. in order to take on the perspective or role of the other. which can then be used to inform successive ones. and to offer support and follow-up information if desired. Like strangers on a train. with its guarantees of confidentiality. If making initial contact is the most uncertain part of the process. at most. to leave at the appropriate time. Early interviews provide the occasion for discovering unanticipated insights. although inter- viewing cannot begin without a specified sample and a well-developed. a few persons. Because depth interviews. Convincing others to contribute to a project that must necessarily be a collective endeavour depends on having a strong belief in the value of the study and a warm but persistent approach. they take place slowly over an extended period of time. in- depth interview more closely resembles the therapeutic interview of clinical practice than the tightly controlled. the sociological interview spans one or two meetings and is not explicitly aimed at helping the respondent. it becomes easier to keep spirits high when the inevitable. unlike surveys. the bounded nature of the interview and the professionally neutral stance of the interviewer make the process of disclosure possible. Unlike therapy. serving as a check and balance for the imperfections of any one interview and allowing flexibility and change. rejec- tion occurs. and they must be constantly re-established with new participants as the study proceeds. Indeed. Indeed. includ- ing sympathy. then con- ducting the interview is the most important and uplifting. both of whom are searching to unravel the mysteries and meanings of a life. however. but the purpose is to enhance full disclosure in order to get to know the person as a unified whole (rather than to retrieve a series of answers from a predetermined set of responses). closed-ended questionnaire used in social surveys. The implicit promise is to listen carefully and supportively. This process creates its own rhythm. Although it may seem paradoxical. the interviewer provides the opportunity for people to step back from their ordinary routines and reflect upon their lives. at least temporarily. In-depth interviews may follow an organized set of ordered ques- tions.210 CHOICES IN CONTEXT batch of letters is sent and a new set of phone calls made. to refrain from drawing judgemental conclusions.9 By creating an impar- tial emotional space. If these initial contacts secure a high level of participa- tion. are conducted by one or. respondents gain the ear of a sympathetic but disinterested person who is not part of their social world. Relationships must be forged (and ended) quickly. if rare. The interview process also demands a willingness to put moral judgement aside. creates a space outside the ‘real’ world in which disclosure and insight can proceed. In-depth interviewing thus depends on creating trust. Its success depends on mustering a range of emotional as well as analytic skills.

there is no perfect interview that can provide the whole story or the real ‘truth’. can offer more than limited insight into general social forces and processes. to examine family processes and trajec- tories (Gerson 2001). OBSERVATION AND INTERVIEWING 211 open-ended. Over time. This surprising interview prompted me to shift from focusing on family structures.11 The interview method necessarily depends on people’s varying abilities to recall the past. which are highlighted in the theo- retical debate over family values. qualitative research depends on being open to the possibil- ity of change. however revealing. When new interviews are more likely to confirm earlier insights than to spark new discoveries. there is also good reason to conclude that a persuasive case can be made to others. While it can be difficult to let early pre- conceptions go. but these experiences signal that the time has arrived to begin wrapping up the data collection process and disengaging from the field. but altered my sampling strategy when an early interview made it clear that single women were theoretically important and just as able to discuss the issues. And. Indeed. For example. comprehend the present and consider the future. . para- digmatic shifts. And it becomes possible to anticipate the answers before the questions are asked. one at a time. There is no reason to conduct a study if the answers are known from the start. some interviews will provide more useful information than others. Inevitably. No single interview. This ‘aha’ experience is especially likely to happen when people offer answers that do not fit the study’s presuppositions or the pre- vailing theoretical debates. Accordingly. I began my study of women’s work and family choices assuming only married women could meaningfully discuss child-bearing decisions. Early in my research on the consequences of changing family forms. When enough interviews have been con- ducted to reach and support theoretical conclusions. it is good news when findings from the empirical world prompt to theoretical and methodological shifts. Collective experiences begin to stand out from interesting but idiosyncratic anecdotes. while others will prompt a new way of seeing concepts and organi- zing principles. each interview will add to the final story. if necessary. in the long run. Only by comparing a series of interviews can the significance of any one of them be fully understood. some interviews will seem unsurprising and unin- teresting. Of course. there is a good chance that theoreti- cal saturation has been reached. general patterns should begin to take shape out of what once appeared to be unique stories. while others find it difficult to recollect past circumstances or contemplate future possibilities. unexpected findings can thus spark theoretical insights that lead to crucial methodological adjustments. I interviewed a young man who described growing up in an outwardly traditional household that actually changed dramati- cally throughout his childhood. By carefully listening to each interview. as the number of interviews grows. Some participants are able to offer great detail and insight. No bell will sound or gun go off. questionnaire. the goal of early interviews is to search for unexpected findings in order to make adjustments and.

are always changing. it is often more useful to be a bit of an outsider (Horowitz 1986). conduct- ing research). I have been able to study people who are different from me in obvious ways. Observation By the time the interviewer enters the field officially. evolve. to some extent. it is time to shift from conducting interviews to concentrating on analyzing them. a researcher needs to present an identity that permits relationships to develop. the observer generally stays in the field for a considerable time period and must thus maintain relationships with people as they move into the future together.212 CHOICES IN CONTEXT Leaving the field can be as difficult as entering it and. my social identity as a person outside the usual social order may have posed a challenge. and sometimes change dramatically over time. it remains an arbitrary decision. which contained gang members. upwardly mobile college-bound youth. in each of my projects. I needed to develop different types of relationships . The social charac- teristics of the observer are often too different to offer any chance to disguise one’s status or purposes. another number to call. the meat of fieldwork and analysis is the long-term observation of relationships and events – some routine. Thus. While introspective accounts add to the picture. There will always be another letter to send. inevitably arises. Yet. but it also provided an opportunity to gain a unique perspective. there also comes a time when more information yields little additional insight and. another person to meet. In the Chicano community. For participant-observers. the participant- observer should be established in the field site. on which all qualitative tech- niques depend. parents of varying outlooks. These relationships are often complex. this strategy is both ethically questionable and frequently impossible. another life to explore. however. Whether as a woman in a man’s world or an older white woman in a world of minority teen mothers. While some have argued that the researcher should disguise that role and simply join the setting. makes it possible to ask probing and even silly or stupid ques- tions. Issues. and will vary among the people and groups within the setting. and political activists. arise in the field that are similar to those faced by an interviewer as each strives to develop relationships of trust and goodwill. nevertheless. to remain in places from which insiders would be excluded. simply makes the data less manageable. the issue of one’s role as a researcher. Fortunately. to violate rules that insiders could not violate. While observations occur in the present. The outsider status of the observer. like that of the interviewer. instead. the research challenge is to focus on events as they unfold and relationships as they evolve. including matters of trust and confidentiality. some not – as they develop.13 Since the fieldworker must spend a substantial period of time in a ‘natu- ral’ setting doing something that is decidedly not natural (that is. To enter a world in which one is not naturally a part. and to cross social lines between groups who may be in conflict or out of contact.12 At this point.

Every field setting will involve some demands on the researcher. my current project on professions has placed me in a setting in which I am considered an expert. involves negotiation and change. Nor can one entirely succeed at consistently expressing the ’right’ view to the ‘right’ group. like any set of relationships. long-term bonds with many of the people in the Chicano community. In my community research. others collapsed. as I invested more time in the research. whether it is to express one’s views. participant-observation involves immersion in a community or organizational setting where people are interacting regularly. They can also interfere with the research process or make leaving difficult. feel confident that the information they provide will not be shared with others. over. Over time. Continual interest in my opinions and advice makes it difficult to ask naive questions that might uncover how others account for particular events or actions. . of necessity. therefore. it becomes increasingly difficult to avoid getting lost in the setting. I became disenchanted with a number of the staff running the social service programme in the teen mothers project. While I developed strong. a sense of relief swept over me. It can also be difficult to know when and how to leave. Indeed. I began to hear ‘confidential’ stories. In contrast. If the interview situation involves discrete encounters. although they can be positive or negative. The gendered norms of the community circumscribed my options and required I spend less time with gang members. participant-observation involves a web of relationships that. to account for one’s presence. Unlike interviews. About six weeks after I began ‘hanging’ regularly with one of the gangs. When the programme ended. for example. and my research was. Strong emo- tional ties are not unusual. At least. It is not possible to remain continually neutral over a period of several years. They must. exit was relatively easy. a central challenge emerged from others’ demands that I become a member of the community and abide by the local rules for women. At about the same time. usually with unconnected individuals. to avoid feeling part of the community. especially since expressing contradictory views will gener- ally become known as people talk among themselves. which are short term and private. and to avoid taking sides. As I left the building each day. These stories were purposely concocted to test my ability to keep a secret. and all needed constant work. personal feelings also evolve. As time in the field unfolds. in this case. I began to feel the same relief expressed by the teens enrolled in the programme. OBSERVATION AND INTERVIEWING 213 with different people. Each of these bonds also changed over the course of more than three years. which I later learned were fictional. as some relationships grew stronger. they provided the means by which those involved in illicit activities were able to accept me as a confidante. I began to wish I could avoid many of the social workers and even remove the programme from their grasp. Although these tests posed dangers. or to take sides. gang members also increased the amount of illegal activities conducted in my presence in order to test my trustworthiness amid rumours that I might be a new type of narcotics agent.

There is rarely a clear or consistent deadline for ending a project or sever- ing the relationships it created. Once the questions are posed in a particular way. and people’s lives continue to change. publish- ing deadline or fortuitous end-of-the-field situation. it is time to leave. These deci- sions require personal discretion and are likely to vary with the field experi- ence. however. While each event may seem new. however. and the logical demands of the project. the sociological interview requires forging a close bond in a short and bounded time period and then letting go. provide much guidance about the future of field relationships. the choice of a research strategy becomes clear. categories and theories. Like the therapeutic process. however. participant-observation is likely to be the method of choice. and negotiating ongoing relationships over a period of years is appealing.214 CHOICES IN CONTEXT A project can end because the field site changes or disappears or because someone else (perhaps a publisher or dissertation supervisor) says it is time to stop. The interviewer. and am I satisfied with my evidence? Can I tell a reasonably com- plete sociological story? When the answer to these questions is ‘yes’. there comes a time when these new occurrences offer little that is new sociologically. Participant-observers must justify and explain their presence to those being studied. Do I have enough evi- dence to persuade audiences? Have I developed my concepts. the decision to leave the field is based on an assessment that enough has been uncovered and the time has come to move to a new stage of analysis. in contrast. I returned to conduct a follow-up prior to writing Honor and the American Dream (Horowitz 1983) and I continue to see some of the people who became friends. The bounded yet intense nature of the in-depth interview creates a different set of contingencies. Their own social experiences and statuses matter. the decision to leave needs to emerge from the research process itself. becoming immersed in the lives of its members. If the prospect of joining a community. Even though I left the field in Chicago in order to finish my degree. the question. Such a pragmatic approach cannot. Yet. They must be prepared to share their views and make personal dis- closures. The process of developing many intense but circumscribed relation- ships draws on different personal resources than does the process of main- taining a small number of multi-stranded relationships over an extended period of time. In the absence of a faculty adviser. needs to offer support and sympathy while also withholding . Most often. one cannot choose a method without making sure that the question to be answered is tightly connected to that method. Conclusion The choice to interview or observe depends on the personal preferences and styles of the researcher. The different nature of the methods also poses different obstacles.14 Personal interests and varying attractions to particular research strategies surely influence the kinds of questions we ask.

but they share the common goal of transforming strangers into confidants. Despite the stylistic and analytic differences. Yet the hard work would not be worth the effort if it . observation takes place in natural social settings. while interviewers must traverse a wide area in order to meet a large number of people. While a degree of withholding is needed during the interview. Similarly. talk and evolving situations. a fuller and more open exchange of personal opinions and experiences can take place when it is over. To discern patterns and links. Observation and interviewing also focus on different levels of analysis and tend to produce different types of findings. group processes. while interviews ask people to take time out from the ‘actually occurring world’ to contemplate from a distance the trajectory of their lives from past to present to imagined future. The challenge is to create an environment in which the participant feels accepted and thus free to dis- close and reflect honestly on controversial personal and political issues. or second- ary. are never clear cut. the success of any qualitative approach requires creating bonds across the researcher–respondent divide. hired interviewers.) A sympathetic and encouraging ear is not only helpful but also necessary. and interviewing can extend well beyond short-term meetings to encompass multiple interviews with the same person over an extended period of time. Kathleen’s interviews have delved into such highly charged experiences as sexual infidelity. observers must spend a long time in a few settings. moreover. unmediated by survey instruments. criminal activity and wife battering. even in the short run. Berger and Luckmann 1967). archival sources. Understanding and encouragement are essential aspects of an interview. for example. but this must be balanced with a certain degree of circumspection. The thin line between insider and outsider status is one that must always be negotiated. ANALYSING FINDINGS AND DEVELOPING THEORY Qualitative techniques offer a way to confront the messy ‘facts’ of social life directly. Interviewing provides a way to uncover the motives. Fieldwork can involve study- ing settings for which the researcher possesses advanced knowledge and familiarity. both methods offer the chance to peer beyond the surface into the inner workings of other social worlds and to see the larger world from others’ perspectives (see Mead 1964 [1934]. (Over the years. Observation necessarily attends to interactions. however. In the end. Both methods typically involve intense physical and emotional immer- sion. needs to be socially and politically neutral so that the respondent does not feel judged or tempted to say only what he or she imagines the interviewer wishes to hear. The distinctions between interviewing and fieldwork. Such feedback. which are embedded in ongoing daily patterns. meanings and conflicts experienced by individuals as they respond to social and interpersonal situa- tions and conflicts. These relationships may emerge in different ways and take different forms. OBSERVATION AND INTERVIEWING 215 information that might influence a respondent’s answers.

and the work remains unfinished until this is accomplished. employ different approaches to analysing data and reaching theoretical conclusions. consolidating new discoveries into innovative explanations would depend on either luck or intuitive genius. brilliant insight need not be accidental. Fortunately. Participant-observation. The purpose of good research design is to create the optimal conditions for making a theoretical contribution. and hard work in the earlier phases reaps immense rewards at the end. not individual characteristics. The first involves a careful reading of all interview transcripts. Interviewing and observation share a set of conceptual and practical tools for arriving at theoretical conclusions. The careful attention to data collection produces material that is reliable and believable. While never relinquishing the inspiration of a sociological imagination. In research as in baseball. Even interviews whose meaning seemed straightforward when conducted can take on a new resonance in the context of all the others.216 CHOICES IN CONTEXT failed to generate new ways of making sense of social arrangements and processes. Interviewing and observation. Certainly. it takes a new and more directed form after all interviews have been collected and transcribed. what are the . ‘Luck is the residue of design’. nevertheless. The time has come to transform a series of individual interviews into a coherent and theoretically trenchant argument about group structures and processes. in con- trast. collectivities and interaction pat- terns. are more likely to form the basic units of analysis. And the vision imposed by one person or. it never hurts to have a bit of both. In the absence of careful forethought. At this point. Carefully designed and implemented research sets the stage for theoretical breakthroughs. generally focuses analysis on groups. generates a focused analysis in a unified. The tight control and per- sonal scale of qualitative projects engenders material appropriate for answering the questions posed. Interactions among individuals. This is the time to ask a set of crucial questions about each respondent and about the group as a whole: what general shape does each person’s life take? What general paths have people followed? Considering the group as a whole. it becomes both possible and neces- sary to step back and seek the shape of the forest amid the trees. These are just some of the dif- ferences in how interviewers and participant-observers may analyse their material in the search for theoretical discoveries. there are nevertheless some identifiable steps to follow in the analytic process. at most a small group of people. which become a lens through which to view social contexts and arrangements. Individual lives are seen to embody larger structural and cultural formations. Interviews focus attention on individual biographies. coherent voice. as Branch Rickey famously argued. Interviewing Although analysis begins the first time an interviewer sits down with a par- ticipant.

Inevitably. and which seem obvious? What makes one interview more interesting than another? By subjecting each interview to close scrutiny. careful reading has created a feel for the interviews as a whole. regardless of the ultimate choice. As I immersed myself in the interviews. a new set of categories emerged. After. this stage involves identifying and measuring the factors and processes that help account for these outcomes. then. but it should become increasingly easier to organize each interview’s facts within a larger analytic scheme. and more compelling than others. and what general form does each group take? Which interviews seem interest- ing. identify the range of factors and processes that may or may not con- tribute to its explanation (the independent variables). In the end. I approach it in ways that resemble the logic of survey research: first. more intensely experienced. all ambiguity cannot be resolved. some cases will provide the purest examples of each type. the interview material can be organized into analytic groupings. I realized that the more interesting interviews revealed personal biographies marked by significant and unexpected change while the less intense interviews involved lives that followed anticipated and relatively predictable paths. it is helpful to ask how and to what extent prevailing theories are either supported or undermined. Once the categories that provide an explanatory focus are clarified. I began to realize that. I began to appreciate the importance of searching for the unexpected in my own data as well. If all goes well. The dimensions of importance and the boundaries between categories become clearer as decisions are made about where each interview belongs in the overall scheme. as I reread and rearranged the interviews into shifting groups. The ambiguous cases point the way toward more sophisticated categories that can do a better job of capturing the complexity of social experience. I set out to compare women who chose to become full-time mothers with those who preferred to work. As I solved this puzzle. and others will seem ambiguous and hard to place. OBSERVATION AND INTERVIEWING 217 range of important outcomes? How does the sample divide into groups. This step must necessarily be more systematic. Which findings suggest that earlier theoretical . In my research on women’s choices. an interactive and iterative process helps define an emerging set of categories or ‘ideal types’ that become the project’s explanatory foci. The study shifted from the analysis of static choices to a focus on pathways of change and stability. If the first task involves clarifying outcome categories. identify the phenomenon to be explained (or the dependent variable) and. these categories will be quite different than the ones that seemed obvious before the study began. and only after. a set of categories for organizing analysis will begin to emerge. it is time to begin the more formal process of creating analytic categories and concepts. Here. Slowly. either by combining it with motherhood or choosing to remain childless. however. By moving back and forth between data and concepts. some biographies were more complicated. and by comparing it to the others. And as I began to look for the unexpected experi- ences and turning points that marked change in the lives of my respondents. Indeed.

analysis and data collection are inseparably intertwined. and connections between categories explored – always with an eye toward where the observer was. When I compared socialization experiences with other factors. These findings also prompted me to focus on the turning points in people’s lives and on the events that trigger change. The long-term shape of social change thus depends on the innovative strategies that people develop in the face of unavoidable but socially con- structed dilemmas. for example. as . Instead of passively accepting the injunction not to work when their children were young. comparative analysis of interview biographies provides a method to clarify the interaction between structure and action. women were more likely to change their child-rearing beliefs in order to justify combining work and motherhood. Men who grew up to be dis- tant from their own children pointed to the example of their own distant fathers as an important cause. in the form of oppor- tunities and constraints. Much of the theory is developed from observations as they are gathered. Such situations illumine the social conditions that encourage or even require change. First. shape adult pathways and strategies and allow people to use their childhood experiences in diverse and innovative ways. It draws attention to the role of institutional arrangements in cre- ating or preventing personal change and also to the role of creative human action in shaping the contours of institutional change. and how it was seen. As my analysis revealed the mal- leability of early childhood experiences. a new framework took shape that moved beyond the prevailing focus on gendered personalities instilled in childhood to focus on how institutional arrangements. Then. but those who became deeply involved in rearing their children pointed to their equally distant fathers as negative influences and examples of what to avoid. In this way. I began to search for other factors to explain why some men become involved fathers while others confined their parental involvement to breadwinning and still others abandoned their off- spring. the categories are constructed and tested in the field. Looking at critical choice points also draws attention to the active and creative ways that people respond to unex- pected circumstances.218 CHOICES IN CONTEXT assumptions are inadequate or incomplete? In what direction do these findings point for a more powerful explanation? In my research on men’s fathering patterns. I discovered that adult opportunities and contexts served to either reinforce or undermine earlier outlooks and orientations. such as work histories and relationships with women. what was seen. Most changed their outlooks over time and then reinterpreted the past in light of newly emerging perspectives. The systematic. and they sug- gest that stability results from the convergence of institutional forces which make change difficult or impossible. I discovered that early childhood models did not explain adult choices. Observation In participant-observation.

Other providers raised problems and issues and asked the teens to resolve them. The same patterns quickly emerged among these new groups. I observed these processes over time as a second session began with a new group of teens and some new staff as well. The major story may not become clear until all the minor themes and concepts have been developed. did success or failure depend on the structure of the programme itself? Studying one programme could not uncover the relationship between time spent on activi- ties or the kinds of services provided and various outcomes. 1998) and Glaser and Strauss (1967) argued. Some told the teens what to do and how to do it. attitudes and behaviour (Becker 1958. it is essential to revise. I noticed that the young mothers became involved and interested in some classes but uninterested and occasionally angry during others. 1998). OBSERVATION AND INTERVIEWING 219 Becker (1958. My notes revealed that social workers treated the teens in different ways and used different interactional strategies. In Teen Mothers (Horowitz 1995). To see if these patterns were organized and recurring or merely a reflec- tion of particular staff and groups of young women. organization or substantive problem rather than to demonstrate the relationship between abstract categories of values. in contrast. Participant-observation is often used to understand a particular place. there typically remains more to do after leaving the field. Projects sometimes begin with the goal of understanding what is going on. While much of the work of discovering ‘what is going on here?’ occurs during the data collection stage. The arbiters defined their job as helping only the most successful to leave their pasts behind. establishing the organized nature of the relationships. suggesting possible choices that the teens may not have envisioned. until they mesh. they stressed that there were right and wrong answers to everything. It also became clear that the staff targeted different types of teens and used their power in different ways. I developed categories of providers. but the analysis cannot end there. This observation led me to focus on how the social service providers interacted with the teens. In addition to the amount of time or the kinds of services provided. As the pro- gramme developed. Like in-depth interviews. I wanted to understand what happened in programmes that delivered services to teen mothers who were welfare recipients. however. From these early insights. including the ‘arbiters’ (who stressed a rigid worldview based on rules and strict hier- archies) and ‘mediators’ (who stressed change based on learning how to face barriers and resolve difficult situations). With a larger sample of providers. retest and revise again in the process of moving back and forth between data and theory. the study could have explored why the . The mediators. and they compared their own ‘successful’ lives with the ‘wasted’ lives of their students. endeavoured to provide the skills that young mothers needed to envision and choose the lives they wished to lead. Prevailing accounts offered mixed observations about why some programmes succeeded and others did not. group. all field notes must be read several times to develop more refined categories and connec- tions among the categories.

active citizens. finally. the way in which arbiters delivered services increased teen dependency and confirmed their identities as failures on the periphery of society. Their interactional patterns in the programme were consistent with the ways they had talked about their experiences and organized their lives outside of it. My findings demonstrate how policies developed at the national level were translated. understand and explain the forms that social organizations take and the paths socially embedded actors follow. Analysing daily. and to develop concepts and categories that could not be discovered from the vantage point of social actors (much as fish remain unaware of the water in which they swim). which argues that welfare permits and increases the depen- dency of the poor. There can be many paths to the same destination. We have also discovered the innovative strategies that emerge as social actors cope with contextual constraints. interpreted and implemented as they made their way to state-level programmes and then to local staff and. however. Despite our contrasting empirical and analytic styles. Qualitative techniques pro- vide a set of analytic tools to discover. Interaction patterns can also be used to contribute to the under- standing of the outcomes of national-level policies. and the left’s critique of the welfare state. Qualitative researchers are able to step outside the normal categories of social experience. While national and state policies were con- sistent with the notion that teen mothers were (and should be treated as) independent agents. these findings provide support for both the conserva- tive perspective. and different qualitative strategies should ultimately yield complementary theoretical insights. the ultimate goal is theoretical insight.) With so few cases.220 CHOICES IN CONTEXT social workers developed such opposing views. The mediators’ stress on teen involvement and decision-making. we have both became attuned to the theoretical importance of processes of change and the critical turning points that marked significant choices and changes in people’s lives. I could only suggest how the providers’ worldviews were reflected in their experiences outside the pro- gramme. to clients. Ironically. By offering an interesting story about what is going on. socially structured conflicts and unexpected . in-depth interviewing using a life course approach could have expanded the range of study questions. however. micro-level interactions has implications for the study of power. and large scale social processes work in daily life. Conclusion Whether the findings take the form of thick description or explanatory analysis. participant-observation can and should provide insights about how organizations. which argues that welfare prevents the poor from demanding more broad-based legal and social rights of citizenship. to view the whole terrain comparatively. policies. (In that event. provides an alternative social policy approach that encourages teen mothers to become independent.

Observational techniques provide a way to collect direct information about how individuals. enrich. Finally. simple dichotomies are misleading and inaccurate. more complete. And participant-observers must inquire about the reasons for actions that others in the setting take for granted as well as actions that take place away from the researcher’s presence. in addition to using empirical findings to advance sociolog- ical understanding. Yet interviewers also observe and use their observations to gather supporting information about the lives of participants. how turning points and triggering situations illumine the organizational or structural factors that normally remain hidden and may produce outcomes that appear preor- dained. in different ways. and ethnographers often find themselves probing for the significance of observed actions and interactions through conversations. both approaches aim to uncover the unexpected. A participant-observer may decide that interviews may illuminate a particular issue. Taken together. and to discover new ways of understanding taken-for-granted or apparently well-understood social arrangements and processes. The pertinent questions always remain: which approach is most appropriate to answer the questions being posed? And how best can the chosen method be carried out? Qualitative researchers should choose the method that best fits their theoretical concerns and personal strengths and then strive to enact it as carefully and thoroughly . Ultimately. We have discovered. If each approach offers its own distinct view of social reality and its own method for discovering the unexpected. It is surely counterproductive to try to assess which method is ‘better’. and more complex view of social life than either can offer on its own. cast doubt on. OBSERVATION AND INTERVIEWING 221 opportunities. Each method thus relies on strategies that are central to the other. to make sense of puzzling or paradoxical phenomena. Interviews offer a systematic way to uncover people’s experiences over time as well as their perceptions. BEYOND DICHOTOMOUS DISTINCTIONS Although we have highlighted the distinctions between interviewing and observation. we also endeavoured to use sociological analysis to understand and benefit human society. both methods provide a richer. careful fieldwork inevitably involves both observation and interviewing. An interviewer may discover in the interviews a place where participant-observation may be useful to add dimension to a particu- lar aspect of the project. and make unexpected discoveries. they often converge in practice. Good interviewers rely on observation to confirm. Each approach permits the researcher to formulate new strategies as the data col- lection develops. Even if the epistemological assump- tions on which each approach is based differ in theory. motives and accounts of these experiences and actions. Some projects explicitly combine the two. groups and organizations behave in a range of settings as well as a way to observe how people explain their behaviour to each other.

Unlike the focus group where the participants respond to each other in addition to the researcher. 8 Quinceaneras are parties with communions given for 15-year-old girls to mark the transition from childhood to young adulthood (Horowitz 1993). As a political statement. 9 The first obligation of the interviewer. argue that a participant-observation study should begin with a specified theoretical problem. during an interview. 6 The terms ‘interview schedule’ and ‘questionnaire’ are used interchangeably to refer to the open-ended. several paths to the same end. (1991). though many of the older residents at the time would have preferred Mexican American. and often are. The ‘best’ method is the one that is well conceived and carefully executed. or a public place. 2 Burawoy et al. Their work will be enriched by eschewing rigid or invidious distinctions and using as many research strategies as needed. Liebow 1967) or a culture of poverty (Miller 1958. the respondent focuses only on the interviewer. is to ‘do no harm’. it meant that those who were Chicano would actively choose between aspects of the United States culture and that of Mexico. for example. NOTES 1 This dialogue has developed not only out of our personal research experiences (in which Gerson relies on in-depth interviewing and Horowitz on participant- observation). Any period longer than that will tire both the interviewer and respondent. which utilized Burgess’s original social areas and provided census statistics and a short history of the area. ACKNOWLEDGEMENTS Our thanks to Amanda Coffey for insightful and helpful comments on an earlier draft of this chapter.222 CHOICES IN CONTEXT as possible. 5 At the time. Lewis 1966). The qualitative findings and analysis will then be rich in theoretical potential. Most interviews take between 2 and 4 hours. there was a prolonged debate about the extent to which poverty was a result of structural inequalities (Gans 1962. It is best to arrange a time and place to meet without interruption . 7 People involved in the underground economy or delinquents not in jail are examples of groups who can be found only by voluntary selection procedures such as snowballing or advertising (Inciardi et al. An instrument of this kind is needed to provide an orderly experience to the respondent and to insure that each person is asked a compara- ble set of questions. the choice should be left to the respondent. an office. like that of the physician. 1993). What matters most is to find a place that affords the time and space to relax. but also in the process of co-teaching a graduate course on the full range of qualitative methods. 10 Homes provide the greatest opportunity to observe people in a naturalistic set- ting. but structured set of questions that are used to guide an in-depth interview. There can be. 3 I found this neighbourhood in the 1970 Chicago Fact Book. Whether the interview location is a home. 4 I use the designation Chicano because the young people preferred it. Privacy is the most important condition for self-disclosure.

Chicago: University of Chicago Press. Horowitz. H. we take all possible measures to minimize both the existence and perception of a power imbalance. and both of us are always mindful of how our position as a researcher might affect our research and analysis. Glaser. (1993) No Man’s Land: Men’s Changing Commitments to Family and Work. (1991) Ethnography Unbound. friends. New York: Basic Books. We agree with Mason that it is important to explore the context surrounding and running through the interviews. (1957) ‘Participant observation and interviewing’. P. and Geer. OBSERVATION AND INTERVIEWING 223 or observation by parents. Marshall. Becker. Berkeley. M. H. change might occur either within the span of a meeting or across a number of meetings. Horowitz. (1998) Tricks of the Trade. We emphasize the importance of using participant- observation as well as seeking patterns both within and across interviews to help contextualize our findings. (1986) ‘Remaining an outsider: membership as a threat to research rapport’. and Strauss. data is collected at many meetings. H. (1983) Honor and the American Dream: Culture and Identity in a Chicano Community. Gerson.W.R. (1962) The Urban Villagers. and Luckmann. American Sociological Review 23: 652–660. (1958) ‘Problems of inference and proof in participant observation’. but it does provide an opportunity to observe interactions and follow up on information gleaned in the interview. lovers or co-workers. The arrival of others is not ideal. Career. (2001) ‘Children of the gender revolution: some theoretical questions and findings from the field’. CA: University of California Press. in V. T. Urban Life 14: 409–430. A. (1985) Hard Choices: How Women Decide about Work. Chicago: Aldine. H. In studies of meetings where decisions are made. R. 13 There are many studies that focus on discrete time units. Burawoy. 14 We are more sanguine than Tracey Reynolds (Chapter 14 in this volume) about the ability of researchers to bridge the power difference between the researcher and the informant. New York: Anchor. and Motherhood. In such a case. Yet neither of us has experienced significant problems gaining access to informants. REFERENCES Becker. Becker. R. B. Such differences should never be ignored in either the data- gathering or the analytic process. Human Organization 16: 28–32. New York: Free Press. . Heinz. B. 1995) discusses these issues. Toronto: University of Toronto Press. for example. Berkeley. Horowitz (1983. (1967) The Discovery of Grounded Theory. NJ: Rutgers University Press. CA: University of California Press. W. 12 In the case of 2–4-hour interviews. Verma (eds) Restructuring Work and the Life Course. Gerson. et al. K. H. K. Krueger and A. New Brunswick. K. Gerson. 11 Both Jennifer Mason and Steph Lawler (Chapters 10 and 11 in this volume) are concerned with the construction and interpretation of the interview. Needless to say. Berger. Gans. my rule of thumb it that fewer than 60 inter- views cannot support convincing conclusions and more than 150 produce too much material to analyse effectively and expeditiously. (1967) The Social Construction of Reality.

(1967) Tally’s Corner. Marriage and Family Review 19: 257–280. (1958) ‘Lower class culture as a generating milieu for gang delinquency’. (1994) ‘Privileging fieldwork over interviews.. and McMahon. New York: Oxford University Press.224 CHOICES IN CONTEXT Horowitz. . Kleinman. (1993) Street Kids. N. R. (1966) La Vida. Lewis. G. S. R. C.B. American Sociological Review 30 (Dec.. Chicago: University of Chicago Press. Symbolic Interaction 17: 37–50. A. M. MA: Little. Belmont. with Stenross. (1964 [1934]) On Social Psychology.W. Ryder. (1965) ‘The cohort as a concept in the study of social change’. (1959) The Sociological Imagination. CA: Wadsworth. Miller. New York: Random House. Street Drugs and Street Crime. and Pottieger. Boston. J. O. W. (1995) Teen Mothers: Citizens or Dependents? Chicago: University of Chicago Press. Inciardi. Mills. (1993) ‘The power of ritual in a Chicano community’. Journal of Social Issues 14: 3–14. Horowitz.H. Mead. R. E. consequences for identity and practice’.): 843–861. Brown. Liebow. B. Horowitz.

what we assume is possible from asking questions and from listening to answers. or subjects. as Silverman has warned. But the elevation of just one humble research method to such heights. pursue and satisfy our theoretical orientations in our research. and the purpose is achieved through active engagement by interviewer and interviewee around relevant issues. are all ways in which we express. Interview methodology begins from the assumption that it is possible to investigate elements of the social by asking people to talk. It has its roots in a range of theoretical and epistemological traditions. and which assume or emphasize the centrality of talk and text in our ways of knowing about the social world. or which layers or elements of ‘the social’ they illuminate. listening and interpreting Jennifer Mason In qualitative research. Good interviewing is clearly in no small part about an inter- viewer’s skills in asking.10 QUALITATIVE INTERVIEWING: Asking. Asking. and to gather or construct knowledge by listening to and interpreting what they say and to how they say it. listening and interpretation are theoretical projects in the sense that how we ask questions. In this chapter I shall examine some of the issues and challenges raised by the use of interview methods in qualitative research. is not without its costs. The style is conversational. flexible and fluid. I shall discuss some of the key questions with which researchers have to deal in the process . all of which give some privilege to the accounts of social actors. topics and experi- ences during the interview itself. agents. This interactive. and what kind of knowledge we hear answers to be. rather than questions about skill and technique in interviewing. but these are more than skills which can simply be acquired and deployed. situational and generative approach to the acquisition of data is usually contrasted with the more struc- tured composition and uniform style of a survey interview. and perhaps most importantly about how well (or badly) they do what they say they do. as data sources. Yet the popularity of interview methods among qualitative researchers is striking. individuals. which I want to focus on here. listening and interpretation (Mason 1996). interviews are usually taken to involve some form of ‘conversation with a purpose’ (Burgess 1984: 102). talk and text constitute. to the point where they are commonly taken to be ‘the gold standard of qualitative research’ (Silverman 2000: 291–292). It is these issues. There is less consensus about what kinds of data such accounts.

and specifically in ideas about how parenting should be done. or oper- ate.226 CHOICES IN CONTEXT of asking. listening and interpretation. Kvale describes something similar and uses the metaphor of ‘interviewer as miner’ (Kvale 1996: 3). negotiated and contextual. This is one type of theoretical project. Finally. Thus interviewing becomes the art of knowledge excavation and the task is to enable the interviewee to give the relevant information in as accu- rate and complete a manner as possible. the researcher may suspect that values and moralities of parenting are less like tangible things existing out there or in there whose nature and shape can be charted. Suppose a researcher is interested in questions about contemporary parent- ing. . but how should a researcher direct or drive the conversation to ensure that the interview gen- erates data appropriate to their research questions or their ‘intellectual puz- zle’ (Mason 1996)? Two key issues to work out here are first. Alternatively. and illustrate how decisions which are made about each of these constitute the theoretical project. and the interview an information-gathering exercise? Or is the relationship of interviewee and interview to knowledge construction more complex than this? Let us take an example to help establish some of the possibilities here. feel or say what seems best or the ‘right thing’ through the way they parent in specific circumstances. where is the social phenomenon or process which is being investigated thought to be located (the location question)? And second. on what basis can the inter- viewee and the interview illuminate it? Is the interviewee a straightforward informant. but it is clearly based on a fairly simplistic onto- logy (theory of what the social is) and epistemology (theory of how the social can be known). morality might be a form of practice such as for example where people try to do. In this version. I shall consider some of the limitations of the interview method’s reliance on talk and text. instead of moralities of good parenting existing in clear and discoverable attitudes. what is good parenting and what is bad. Thus. This orientation to the loca- tion question casts the interviewee as informant – on the social world. or on themselves – and the job of the interview is to unearth the relevant informa- tion. negotiations and people’s contextual reasoning processes – in other words if they are the processes – then asking about ‘good parenting’ in a decontextual way is unlikely to produce meaningful data. only through practices. ideas and values have a tangible and static existence on particular planes of the social. EXCAVATION OR CONSTRUCTION? Interviews involve dialogue between two or more people.1 One possible orientation to the ‘location question’ is that ideas and values about parenting exist ‘out there’ – in the social world in the form perhaps of ideologies – or ‘in there’ – in people’s attitudes and beliefs. If moralities take shape. and instead they are processes of social construction and practices which are fluid.

which in any case may not be observable in the conventional sense. One way is to ask the interviewee to recount or narrate relevant situations. It cannot unearth the relevant data. then the task is to work out how to organize the asking and the listening so as to create the best conditions for the construction of meaningful knowledge (about moralities of parenting. One way to attempt to resolve this dilemma is to treat the interview as a site of knowledge construction. This. ontology and epistemology. questions should be couched in specifics rather than generalities. is a different type of theoretical project and one which I would argue is based on a more sophisticated. act and interpret (see also Holstein and Gubrium 1995). and more satisfactory. Yet the interview. or whatever). The assumption here is that by grounding the interview dialogue in relevant contexts. the researcher needs to devise questions and modes of asking which both anticipate and discover the range of contexts in which moralities of parent- ing get done by or in relation to their interviewee. might focus upon the detail of how they ‘do parenting’ on an everyday basis or at ‘definitive moments’ by asking. or the interviewer cannot be in all of the relevant contexts to witness the operation of practices and processes. therefore. Questions may have a biographical or life story orientation. particularly if the researcher suspects that moralities of parenting are processual over time and lifetimes. then. GENERALITIES OR SPECIFICS? If interviewing – and asking. and the interviewee and interviewer as co-participants in the process. because the phenomenon under research does not have a static decontextual and therefore uncoverable existence. knowledge about moralities practised outside the interview setting can be constructed on the basis of interactions within it because the operation of morality as process or practice becomes more possible to articulate through the specifics of the narrative. how they organize various aspects of their relationships with their children and what matters to them (and conversely what does not) in the different contexts raised. Knowledge gained in this way is a co-production since it is dependent upon the combined efforts of interviewer and interviewee in conjuring up the rele- vant contexts from which they think. listening and interpretation as theoretical pro- jects – is the art of construction rather than excavation. Questions. using the interviewee as informant. then. and is based on the idea . QUALITATIVE INTERVIEWING 227 This means that the interview method is up against some major challenges. contexts and events so they can effectively con- struct or reconstruct (to continue with the same example) moralities of parenting in the interview setting. for example. this means that instead of starting from interview ques- tions which invite generalities or abstractions like ‘What is good parenting?’. talk. The practice of asking about the everyday has a long established pedigree in qualitative research. In practical terms. In this type of theoretical project.

the practice of focusing on definitive moments or points of renegotiation or change is well established. that is in the interpretation. Mason 2000). but we would not have the tools or materials to fashion this into a generalization of our own about how that related to other contexts. then surely the very fact that they dealt with the question and devised an answer must mean something? The difficulty lies in working out what it does mean. my own research into family and kinship suggests that inter- viewees often ask for clarification of abstract and generalized questions because these kinds of questions do not make immediate sense and people find it difficult to formulate an answer (see especially Finch and Mason 1993. if they answer the question by relating it to their own or someone else’s circumstances. if they do. given that it is ultimately based on the same assumption that morali- ties are contextual. This is a problem. Where they do.228 CHOICES IN CONTEXT that the way people make sense of the social is grounded in their everyday – even routine – experiences. because if further clarification and possibly contextualization is required for the question to make sense and for an answer to seem possible. then it seems likely that there is no level of the social which corresponds to the abstract version of the question. 2000. an interviewee is prepared to say. for example. Hollway and Jefferson (2000) report similar difficulties in their research into fear of crime. and is much inferior to the strategy of beginning with contexts relevant to the inter- viewee. It may. But does this mean that it is never useful or productive to ask interviewees for general- ities or abstractions? It may of course be interesting to see whether people are prepared to answer these kinds of questions. that ‘A good parent is one who is caring but also exercises discipline’. this way of situating inter- viewees’ responses is likely to be a rather hit-and-miss affair. However. however. how are these connections between the abstract and the situated made? This might yield knowledge about rele- vant contexts and forms of morality. in response to the ‘What is a good parent?’ question. In fact. the answers often appear very cliched and empty of any grounded meaning. That an interviewee would make a statement like this in an interview context might tell us something about the sense of a moral self which they were creating in that setting. Similarly. and that the theoretical project is flawed. If we assume that moralities are processes and practices. can specific rather than abstract questions really illuminate this? One response to this question is to ask whether interviews can ever tell us . Do people never look outside the parochial bound- aries of their own situation to more abstract notions of right and wrong and. then an abstracted or generalized answer cannot make any sense without some knowledge of how it relates to the individual’s practices and experi- ences. since individuals do not inhabit abstract and generalized social worlds (even when they are being interviewed). Of course this raises the question of whether everything about moralities of parenting is contextual. be useful to track the kinds of clarifications that interviewees seek – what kinds of contextual or other information do people require in order to formulate a response? Or. If.

QUALITATIVE INTERVIEWING 229 everything we want to know about our research problem – in this case moralities of parenting – and that is an issue to which I shall return shortly. We derived ‘composite’ moral narratives from our inter- viewees’ many personal stories. specific questions about people’s own experiences can make a much better job of enabling us to analyse whether and how people use abstractions (in this case abstract moral principles of right and wrong) in their practices. than can abstract questions themselves. We explain the logic of this as follows: We have used the concept of narrative as a methodological and analytical device to illustrate some of the consistencies in our interview data but. using a moral discourse. Take the following example of an excerpt from one of my recent research projects on the topic of how families handle inheritance. I still think that it would be good for my children to make their own way. most notably. more than that. in their abstract and their particular guises. to communicate accounts and scenarios which people recognise and. In this sense. and from what they told us of their hopes and fears about inheritance. this. and probably better for them in some way that they should make their own way. good parenting involves not passing on too much in a material sense to one’s children. which they fear. They shouldn’t be helped too much. … It sounds a bit reac- tionary. In analysing the very many moral accounts which that research project generated from its specific and contextual questions. I think that they shouldn’t have things too easily … I don’t really assume that I’ve got to hand on the wealth that we will have to the children … it’s probably better for me if I use some of it before I go. The narratives we have used tend to be expressions of what people think should . we are able to get a better sense of how moralities of parenting are used or operate. He articulates that very clearly. without being asked ‘What is good parenting?’ His account blends the general (they shouldn’t have things too easily) with the particu- lar (‘good for my children to make their own way’. is conveying a great deal about moralities of parenting. the narrative is an expression of people’s attempts to connect up their own experiences and reasoning with something which they per- ceive to be more generalised. (Finch and Mason 2000: 125–126) For this interviewee. but what he says is a response to questions about his specific circumstances – in this case as a member of the first generation in his own family to be a home- owner – and about his own bequeathing intentions. But the other response is that yes. and the significant point is that many people do this. ‘I don’t really assume that I’ve got to hand on the wealth that we will have to the children’) and is articulated through – because it is located within – a particular context or set of circumstances. like many others.2 This interviewee. But it does build your character. In writ- ing our analysis from that particular project. Janet Finch and I developed the analytical and methodological device of ‘the narrative’ in a rather specific way to help us to engage with these relationships between the general and the particular. than had we simply asked about the rights and wrongs of inheritance and parenting in abstract ways. shouldn’t expect too much.

Smart and Neale 1999). is an important element in the way in which people do kinship and inheritance. this is because . (Finch and Mason 2000: 165. if used flexibly in qualitative inter- views. Also.230 CHOICES IN CONTEXT not happen. However. but are given contextual information – albeit hypothetical – with which to do this. This means that the construction of generalising narra- tives. with the question of what the data might mean. it would be possible to use composite narratives of the kind discussed above). They do not describe what people generally do. because in theory in word it allows the interviewee to discuss their own moral reasoning without having to (publicly) locate what they say in the detail of what may be difficult or private experiences. This may be a particularly useful way of asking questions in ethically and morally sensitive situations. or what they do not want for their own families. as a way of contextualising their own practice. possible on the basis of existing interview data (for example. The vignette is presented to the interviewee. and they could not therefore be used as knowledge about the interviewee’s own situated moral practices. Another more practical reason is that once an interview takes on that kind of abstract discursive style it may be difficult to regain the contextuality so central to the construction of situated knowledge. who then is asked to say what the people involved in the scenario should do. or their own situated moral reasoning. nor do they repre- sent moral rules about what they should do. one good reason to avoid questions which seek or encourage generalities and abstractions is that the knowledge which we produce from these may not be quite what we think it is. In part. The logic is that interviewees are asked to do moral reasoning on the spot. original emphasis) One way to try to shortcut this process is to ask about generalities and abstractions in the first place rather than to derive them from interviewees’ accounts. unless the vignettes had been used as vehicles to get the interview to produce dialogue about these directly. This means that the stories which the narratives contain do not in and of themselves represent an empirical and generalisable real- ity of kinship. but to do it in a situated way through the use of vignettes – hypo- thetical scenarios concocted by the researcher in advance of the interviews. vignettes can allow the researcher to track in a much more contextual and sensitive way any further clarifications which interviewees require. We are left. This seems a more satisfactory way of asking abstract questions because although not directly situated in the interviewees’ own experiences. and are scenarios which they actively try to avoid. however. To sum up. Their connection to contexts other than the interview itself may be tangential at best. it does use notions of situatedness and context in strategic ways as part of a theo- retical project. or what they themselves would do in that situation (Finch 1987). and hence what contextual knowledge is relevant to the reasoning process (for examples of this see Finch and Mason 1993. and what it is may be quite limited when seen in the context of the overall theoretical project we are pur- suing. we learn about people’s practices and moral reasoning through them because many people use narratives like those we have sketched out to locate and make sense of their own (usually oppositional) practices.

among other things. therefore. if an interviewer asks whether an interviewee’s father treats her and her sister equally and she answers that yes. undoubtedly help to shape the nature of the knowledge produced. because despite the use of the term ‘unstructured interview’ in some methodological discourses. we cannot assume on the basis of this answer that the concept of equality itself figures in that interviewee’s family practices. STRUCTURED OR ‘STRUCTURE-FREE’? The types of questions an interviewer asks. experiences and reasoning. For example. I shall expand on this point in the next section. by definition lacks the flexibility and sensitivity to context and particularity required if we are to listen to our interviewees’ ways of interpreting and experiencing the social world. It seems sensible. and the answers once again depend upon their theoretical orientations. to avoid abstractions and generalities in qualita- tive interviews unless we are very sure we have a use for them. it is not possible to conduct a structure-free interview not least because the agendas and assumptions of both interviewer and interviewee will inevitably impose frameworks for meaningful interaction. However. The idea that interviewees may be ‘answering’ questions other than those we are asking them. and making sense of the social world in ways we had not thought of. underpins much of the ‘qualitative’ critique of structured survey interview methods. the structuring principle – interviewees’ own life story – is considered to be . Most qualita- tive researchers try to structure interviews in ways which are meaningful to interviewees (and relevant to the research). A structure or sequence of questions which is rigid. in interpreting data. However. QUALITATIVE INTERVIEWING 231 the construction and reconstruction of relevant contexts in an interview is probably only possible in a sustained way – it is not an activity which can be dipped into and out of. lies behind many qualitative interview strategies. One example of this is life history or biographical interviewing. and to their ways of understanding. and many try to minimize their own role in the process of structuring and in the sequencing of the dialogue. he loves them both. structure an interview. Indeed. and to argue that they are would be to cast the role of the researcher too deter- ministically. but also about the structure or framework for the dialogue. it is very important for researchers to see that sometimes what an interviewee says is not the straightforward answer to the interviewer’s question that it is pre- sumed to be. The problem is not only about how questions are asked (for example in abstract or specific terms). this raises a problem. and how far to. The logic that we should be receptive to what interviewees say. interviews are not just about the asking of questions and the proffering of answers to those questions. and which is devised in advance by the interviewer. and the way they listen to and interpret the answers they are given. The question to be addressed by the researcher is instead how to. Here.

and the sequence is taken from that story. rather than a neutral medium for the simple excavation of facts (Chamberlain and Thompson 1997). Instead. Many life history interviewers will try to facilitate the telling of the story more than to direct it. a purpose or a plot. Hollway and Jefferson (2000). Narratives encourage the construction of a begin- ning. From this per- spective. a sense of progression. unconscious dynamics are a product of attempts to avoid or master anxiety. they are a problem. yet these – in all their messi- ness – are the objects of study. the structural tendencies imposed by the narrative form are of interest in themselves and thus are far from being a problem. sequentiality. provide the key to a person’s Gestalt. they advocate the ‘free association narra- tive’. the associations follow pathways defined by emo- tional motivations.232 CHOICES IN CONTEXT meaningful to them. This method is based on psychoanalytic principles. and an author. The narrative form shapes what is being told in certain ways. because they tidy up and sanitize what are often messy social processes and experiences. and that the way to achieve this is to ask inter- viewees to structure their own life story narrative. leads the researcher to miss highly significant and less formally structured ele- ments of social experience. Thus although this approach may feel more ‘hands off’ from the interviewer’s point of view. but accord- ing to unconscious logic. educa- tional biography) and to follow their own cues. including the identity investments these give rise to. the imperative toward a particular kind of structuring on the interviewee’s part may nevertheless be quite strong. and in particular on the idea that an interview should find ways to tap those elements of the subject’s experience which are not recountable or explainable by themselves – elements which are not authored. for example. that is. According to psychoanalysis. the psychoanalyst is eliciting the kind of narrative that is not structured according to conscious logic. a middle and an end. work biography. experience and perspectives (Miller 2000). For others. the object of study. providing coherence. By eliciting a . and finding ways to elicit narratives from interviewees is a crucial part of that. rather than rational intentions. This suggests that anxieties and attempts to defend against them. argue that a preoccupation with the biographical or other coherence of the narrative. the analysis of narra- tive and discursive conventions and their relationship to social practices and experiences is the theoretical project. family biography. like Chamberlain and Thompson (1997). and the emphasis this places on the capacity of the narrator to know and tell their story. however. The point is that the structure offered – the telling of a life – allows inter- viewees to craft their own narrative around their own concerns. rather than from a pre-devised list of questions. They argue that: by asking the patient to say whatever comes to mind. For some. it is important to appreciate that the narrative itself is a cultural form or genre with its own structural conven- tions. This approach is based on the logic that the significance of social experience will be revealed through contextual data. Even so. sometimes according to specified principles (for example.

contradictions. The irony. CHALLENGING OUR THEORETICAL ORIENTATIONS? My argument so far – that how we ask and listen are theoretical enactments of our assumptions around where the phenomena we are interested in are located. and how the interviewee and interview can illuminate the issues – has rested on the assumption that we know what it is we are looking for. Instead. The Gestalt in which Hollway and Jefferson are interested is not a consciously constructed life narrative. While a common concern of both approaches is to elicit detail. for example through allowing inter- viewees to develop points and stories in depth and return to them at will even though their relevance to the substantive concerns of the research may not be evident. avoidances) and accord them due significance. My critique of general and abstract interview questions is based on the argument not only that these direct attention to wrong or . of course. indeed Hollway and Jefferson developed seven standard starter questions around which they encouraged the development of the narratives. a psychosocial subject. This kind of theoretical project demands that the interview ‘structure’ allows spaces for free association. as does that of Chamberlain and Thompson (1997). Hollway and Jefferson achieved this by train- ing themselves to be the ‘almost invisible. and certainly through not enforcing a particular sequence of questions. be that contextual moral practices. which is not consciously authored and cannot be articulated in conventional narrative form. facilitating catalyst to their [inter- viewees’] stories. schisms and defence mechanisms within the narrative. elisions. whether or not the interviewee is aware of them. QUALITATIVE INTERVIEWING 233 narrative structured according to the principles of free association. Nevertheless Hollway and Jefferson’s work demonstrates par excellence. we secure access to a person’s concerns which would probably not be visible using a more traditional method. therefore.. It demands an act of faith (in psychoanalytic principles) to believe that it is the unconscious rather than something more social or cultural which consequently does the structuring of free association. nor is it an understanding of the place and use of narratives in social practice. for Hollway and Jefferson (2000). it is a psychosocial subject.. narrative analysis has a preoccupation with coherence which we do not share. what is more often unrecognized: that how a researcher deals with issues of interview structure and sequence is always part of their theoretical project. means not imposing a structure on the narrative’ (Hollway and Jefferson 2000: 36). Being “almost invisible” . is that even free-association narratives require some kind of structure to make the telling of them possible. Free associations defy narrative conventions and enable the analyst to pick up on incoherences (for example. or the use and oper- ation of narratives. (Hollway and Jefferson 2000: 37) Allowing interviewees to ‘free associate’. whether or not they acknowledge it. gives researchers using psychoanalytic principles a way into the uncon- scious because it allows them to spot and make sense of connections.

and as a lens for understanding social change and the social world more generally (Giddens 1992. Giddens 1991). and to see things we are not looking for.234 CHOICES IN CONTEXT ‘non-existent’ locations (for example. whether or not this is articulated. continue to hold any sway in what is seen as the new. this means that the capacities for individual action. should they. Sevenhuijsen 1998. fragmented and fluid social order. as forms of agency practised by indi- viduals and in relationships and interactions (rather than as truths dictated from on high) take centre stage (Bauman 1993. the intimate relationships people have with one another are therefore no longer seen as side issues in social science. but we need to explore what is involved to work out how this might feasibly be done. then how do people work out what is right and what is wrong in their relationships with others (if they do not simply follow moral rules and codes)? . the retheorization and exploration of morality and ethics. I take the view that all research has some kind of theoretical orientation. choice and reflexivity are increased while simultaneously people develop a lack of trust in familiar institutions and universal truths (see e. Finch and Mason 1993. abstract attitudes and values). For some. we need to establish the social-theoretical context which makes possible the formulation I have offered that moralities of parenting are contextual. First. agency. The apparent lack of a coherent and uniform moral order or clear sets of rules about what is right and what is wrong are seen as part of far-reaching social changes involving the demise of social structures. For others. Perhaps for the first time in the history of sociology. Smart and Neale 1999). A form of ethics derived from the ‘concrete’ rather than the ‘generalized other’ underpins this (Benhabib 1992: 9–10). 2000. the question of how morality works in personal relationships has begun to occupy the imaginations of many researchers. as do all forms of asking. but also that they miss the point about morality in that they assume it is a thing rather than a process or practice. and the changes we consider to be taking place in that world. If it cannot be said simply that people follow the rules created for them by religious and other institutions.g. The version of morality which I have set out has its own set of theoretical underpinnings. but are considered absolutely central to these new forms of agency and practice. listening and inter- pretation. Beck and Beck- Gernsheim 1995). The ‘postmodern turn’ in social theory has helped to throw issues about morality into the limelight because it questions the degree to which moral absolutes. But how far should we use our interviews to view the social world in different ways. traditions and institutions as organizing principles of the social world. Could they. as well as those we are? Let us continue to explore this question in relation to my moralities of parenting example. ontologi- cally speaking. As a consequence of these changes in the way we theorize the social world. be challenged or tested through interview methodology? My answer is a qualified yes. and therefore there is always some sense in which researchers know what they are looking for. created by higher-order experts and institutions.

to enable them to devise ways of test- ing their own approach both within and beyond interview methodology. for example. and a concern with the agent. however. Another sees human agency not so much as overstated. The ‘discursive turn’ in social theory has claimed the death of the subject. or ‘the proper thing to do’ of any concern or consequence in personal relationships? What forms do they take? How are they expressed or prac- tised? How and when do they change? The language used to frame these kinds of questions – moral practices. But there are other ways of looking at these issues. yet how morality works or. overcoherent story which again has the effect of overplaying agency and rationality? As well as ques- tioning the form which morality takes. rules. we might also question the idea of morality itself. in and of itself. To return to the point about seeing what we are not looking for. self-governing subject who can account for their practices and reveal the logic of those practices in a research interview. but as missing the point altogether. what theoretical baggage does the concept of ‘parenting’ carry – with its implication that this is a skill or a project practised on or ‘done-to’ children. the concern is with multiple subject positions created through. namely that morality is a frame- work. QUALITATIVE INTERVIEWING 235 Have people dispensed with morality altogether? Are questions of right and wrong. . that might feel like rather a meaningless task. Agency. the key question is. in the ways I have suggested. moral agency – itself stands in marked contrast to earlier (even interaction- ist) social scientific concerns with moral rules. traditions or discourses? Does the orientation to moral- ities of parenting and to the process of construction in the interview setting overemphasize the individual. or a set of rules. Is the concept of morality the right one? Does morality have anything to do with parenting? Or is parenting more about practical (and impractical) actions. Instead. detraditionalization and fragmentation plus the potential for human agency and interaction have been overstated. and especially of the rational. The point here is. power relations and so on? Similarly. One has already been alluded to throughout. more accurately. to advocate that researchers develop enough of a sense of alternative conceptualizations of their research problem. could we see moral norms. It also seems highly consistent with an actor/agent-centred form of social inquiry such as the qualitative interview. and is the antithesis. then: if we are looking for moral practices and agency. might very well be. and the centrality of text. norms and codes. language and practice in those discursive processes. According to this approach. and the different types of theoretical project that they might involve. moral discourses. unitary. how morality is constructed discursively. rather than a two (or more) directional relationship extending beyond childhood? Clearly. are not part of the epistemological vocabulary here. and human agency? Does an emphasis on interviewees’ narratives produce an oversanitized. norms and traditions. and proposes a project of ‘decentring the subject’. it is possible to generate a long list of alternative conceptual and theoretical orientations and.


So, for example, to try to avoid an overemphasis on the individual and on
agency, we might consider gaining other forms of data which are less reliant
on the mechanism of self-report, and which reflect on other levels or dimen-
sions of the social (such as observation, demographic data analysis). We
might seek out inconsistencies just as much as consistencies in our interview
questions to try to avoid oversanitizing our data. We might focus less on our
key concepts and more on wider or looser ones – even oppositional ones –
which give us the possibility of seeing other things as well. In the example I
have been using that might involve seeking out ‘what matters’ in
parent–child relationships rather than the narrower concept of morality
specifically. Finally, we might focus on relationships between people, with-
out presupposing anything directional about these, rather than treating
‘parenting’ as a practice done to children, or a set of skills possessed by
parents. This inevitably raises the question of who has the knowledge, the
experience, the defended self, or whatever, that we are interested in. If we
see our focus as relationships (parent–child) rather than individualized prac-
tices or skills possessed (parenting), then parents’ perspectives can provide
data on only part of this. We need to interview children too, at the very least.
In the process of challenging our theoretical orientations, we need also to
ask a more fundamental question about interview methodology, and that is,
are talk and text enough? Do interviews give too much epistemological privi-
lege to the idea of the individual, articulate, rational actor? Can interviews,
whether with carefully selected ranges of individuals, or with groups, ever
tell us about those elements of the social which appear to go beyond or oper-
ate outside individuals – whether these be discourses, or institutions, or sys-
tems? Can they tell us about elements of the social which are not accessible
through talk? While we cannot expect interviews to be able to do everything
for us, I want to conclude with a consideration of some of the limitations
imposed by an emphasis on talk and text.


So far in my discussion I have argued that the way we ask questions, listen
to and interpret talk, all constitute theoretical projects. Underpinning the
kind of theoretical project I have outlined is the assumption that talk and text
are central in our ways of knowing the social world. In qualitative inter-
viewing, ‘talk’ means interactive talk, and some of the enthusiasm for the
method which has emerged in recent years is undoubtedly a reaction against
the asking of questions in less interactive ways, for example through postal
questionnaires and structured questionnaire surveys. Many qualitative
researchers would probably agree with Fontana and Frey (1998) that

as long as many researchers continue to treat respondents as unimportant, faceless
individuals whose only contribution is to fill one more boxed response, the
answers we, as researchers will get will be commensurable with the questions we


ask and with the way we ask them. … The question must be asked person-to-person
if we want it to be answered fully. (Fontana and Frey 1998: 73)

Even Hollway and Jefferson’s (2000) psychosocial subject and intersubjec-
tivity, which operate in part through the unconscious and through feelings
(as well as through talk and conscious interaction, discourses and so on) can
best be known according to them through interview talk, albeit they propose
a focus on the spaces and schisms within dialogue. Although what they are
interested in operates at a level or in ways which cannot be reasoned or
explained (they argue that people are not ‘their own best explainers’), they
nevertheless argue that we can construct knowledge of it by listening to
people’s free associations, connected and disconnected narratives, and inter-
preting them through a psychoanalytic frame. This does of course raise some
political problems around the issue of claims to truth.
The privileging of talk is understandable in the climate described by
Fontana and Frey (1998), yet we should not allow our enthusiasm for the rich
and fascinating data which can be generated in interviews to stop us seeing
some of the limitations of using talk and text to construct knowledge, and to
contemplate ways of overcoming these.
Criticisms of interview and biographical methods have for a long time
pointed to the vagaries of memory, selectivity and deception in interviewees’
accounts (see Chamberlayne et al. 2000 for a useful review) and also to issues
around fluency and divergent linguistic codes. Furthermore, it is important
to engage with the ‘politics of talk’, and to recognize that what counts as lan-
guage, who uses it, what is its nature, what it can mean and do, are not
merely part of a neutral and given reality, but are products of power rela-
tions and struggles (O’Brien and Harris 1999). All of these factors indicate,
with different implications, that we should not read interviewees’ accounts
as straightforward descriptions of social experience.
But the points I want to make move beyond the question of whether or not
we can take interviewees’ accounts at face value, towards questions about
those elements of the social which cannot be expressed through talk, and
which are not situated in talk. Of course the idea that individuals cannot
express everything in which we might be interested in words has long
underpinned observational methods, but there may be elements of social
experience which cannot readily be observed either.
I am referring to processes of thought, feeling, emotion, sentiment and so
on. These may be rather significant and, certainly, one can readily see that
my example of moralities of parenting may be closely bound up with these
or, to put it another way, such moralities may be played out or practised in
those unspoken social locations. There is a danger first that an emphasis on
what can be articulated verbally obscures these and, second, that the ‘dis-
cursive turn’ in social science, with its emphasis on text and the discursive
construction of subject positions, rules these out of the frame altogether.
Yet researching these elements of the unspoken is inherently problematic,
and we may need to find ways of encouraging non-verbal expression to


explore dimensions which people find difficult or impossible to express in
words because, to reiterate an earlier point, we need to create the best contex-
tual conditions for the construction of meaningful knowledge. Methods which
encourage non-verbal expression are increasingly deployed in research with
children. For example, play and drawing are now commonly used as research
tools with children, because it is recognized that they may find certain ideas
and experiences difficult to express in words, and also because key elements
of their social experiences, practices and relations may not occur or be mani-
fest in dialogic form in their everyday lives in the first place. Yet it is strange,
as well as patronizing to children, that such considerations are so rarely
applied to adults also (see Solberg 1996). Furthermore, even in research with
children there is probably some way to go in working out how non-verbal
products and processes should be interpreted and expressed.
It is of course possible to treat non-verbal products and processes as texts
which represent unspoken dimensions of the social world or through which
those dimensions are constituted. This involves a generous definition of ‘text’
but as Devine and Heath point out in their discussion of postmodernism and
empirical research, ‘a text can be anything from a literary text, an official
document, or an interview transcript through to a photograph, a movie
or a building’ (Devine and Heath 1999: 207). Yet constituting non-verbal
products and processes as texts may miss the point about what they are and
what they are meant to be, and may obscure their processual, agentic and
non-discursive nature.
That suggestion is supported by experience of family and kinship
research, where fairly frequent use is made of various different ways of dia-
grammatically mapping and charting the ‘closeness’ or the ‘supportiveness’
of people’s relationships, but where what seems most important is not the
chart or family tree which is produced – the ‘text’ – so much as the some-
times agonized processes which people are observed to go through in trying
to decide which relative or friend goes where in it (Finch and Mason 1993;
Flowerdew 1999). It is interesting that the act of placing a relative in a chart
often is treated by the person doing it as highly significant in itself: some-
thing is being done more than said, and something non-verbal is being
expressed. Incorporating this kind of activity into an interview thus helps to
create a rather different context for the construction of a more non-verbal
kind of knowledge.
Using non- (or semi-) verbal techniques such as these, or photography or
video recording, which consciously and conscientiously move beyond a pre-
occupation with talk or with text, is clearly an important way in which we
might explore non-verbal elements of the social. Although they are often
accompanied by talk, and sometimes used within an interview context, these
methods draw heavily on observational techniques developed by social
anthropologists and ethnographers. We must, however, be mindful of the
epistemological assumptions we make when we employ this kind of methodo-
logy, and when we attempt to interpret our data. In particular, we should not
assume that visual methods, for example, produce knowledge which is


somehow less constructed or more directly representational than verbal
interview methods. The critique of the idea of excavation of data which
I outlined earlier applies with equal force in relation to visual artefacts
and products. Instead the point is to evoke or construct knowledge about
non-discursive experience, and the argument is that using non-verbal
methods should help to create conditions appropriate for the generation of
such knowledge.
Examining the role of non-verbal elements in social relations, and of
objects and artefacts – again not just as texts or representations – is impor-
tant here. To explore moralities of parenting, for example, one might use
people’s personal photographs, objects and possessions as starting points
for discussion and observation, as well as for analysis in their own right. In
a recent study of inheritance, my colleagues and I looked at (often literally)
objects people had inherited and investigated what kinds of objects they
were, where people kept them, considered what their role was in kin rela-
tionships, what they symbolized or expressed, what was their ‘lifespan’
and the changing nature of their ownership, and so on (see Finch and
Mason 2000).
Methods which seek to explore the non-verbal non-verbally can of course
be complemented by more traditional approaches, such as making infer-
ences about, for example, emotions and feelings on the basis of what people
say, how they say it, and what they do not say. We can sometimes discern
whether someone is used to thinking in a particular way by what they say,
and we can infer that something is taken for granted if they do not see it
necessary to mention it. We can sometimes infer how emotionally engaged
they are with a particular issue by how they talk about it, and we can get a
sense of what matters emotionally by observing their demeanour as they
speak. We can ask people to try to articulate elements of the non-verbal, to
tell us how they feel, and what makes them angry or sad.
It is clear that there is a great deal of scope for developing methods which
loosen the grip of talk and text on our research imaginations, and this might
be done under the auspices of a range of different types of theoretical pro-
ject, within and outside of interview methods. At the very least, researchers
using interview methodology should consider carefully whether strategies
which prioritize talk and text can deliver the required goods.


1 Investigation into ‘moralities of parenting’ is one of the key interests of the ESRC
Research Group for the study of Care, Values and the Future of Welfare (CAVA)
at Leeds University, of which I am a member. Further information is available
from the website: www.leeds.ac.uk/cava
2 This project, ‘Inheritance, Property and Family Relationships’, was funded by
ESRC, grant no. R000232035. It was directed by Janet Finch, Jennifer Mason and
Judith Masson, and the research officers were Lynn Hayes and Lorraine Wallis.



Bauman, Z. (1993) Postmodern Ethics. Oxford: Blackwell.
Beck, U. and Beck-Gernsheim, E. (1995) The Normal Chaos of Love. Cambridge:
Benhabib, S. (1992) Situating the Self: Gender, Community and Postmodernism in
Contemporary Ethics. Cambridge: Polity.
Burgess, R.G. (1984) In the Field: An Introduction to Field Research. London: Allen &
Chamberlain, M. and Thompson, P. (1997) ‘Genre and narrative in life stories’, in
M. Chamberlain and P. Thompson (eds) Narrative and Genre. London:
Chamberlayne, P., Bornat, J. and Wengraf, T. (2000) The Turn to Biographical
Methods in Social Science. London: Routledge.
Devine, F. and Heath, S. (1999) Sociological Research Methods in Context. London:
Finch, J. (1987) ‘The vignette technique in survey research’, Sociology
21(1): 105–114.
Finch, J. and Mason, J. (1993) Negotiating Family Responsibilities. London:
Finch, J. and Mason, J. (2000) Passing On: Kinship and Inheritance in England.
London: Routledge.
Flowerdew, J. (1999) ‘Reformulating familiar concerns: parents in stepfamilies’,
unpublished PhD thesis, University of Leeds.
Fontana, A. and Frey, J.H. (1998) ‘Interviewing: the art of science’, in N.K. Denzin
and Y.S. Lincoln (eds) Collecting and Interpreting Qualitative Materials. London:
Giddens, A. (1991) Modernity and Self-Identity: Self and Society in the Late Modern
Age. Cambridge: Polity.
Giddens, A. (1992) The Transformation of Intimacy: Sexuality, Love and Eroticism in
Modern Societies. Cambridge: Polity.
Hollway, W. and Jefferson, T. (2000) Doing Qualitative Research Differently: Free
Association, Narrative and the Interview Method. London: Sage.
Holstein, J. and Gubrium, J. (1995) The Active Interview. Thousand Oaks, CA: Sage.
Kvale, S. (1996) InterViews: An Introduction to Qualitative Research Interviewing.
London: Sage.
Mason, J. (1996) Qualitative Researching. London: Sage.
Mason, J. (2000) ‘Deciding where to live: relational reasoning and narratives of the
self’, working paper no. 19, Centre for Research on Family, Kinship and
Childhood, University of Leeds.
Miller, R.L. (2000) Researching Life Stories and Family Histories. London: Sage.
O’Brien, M. and Harris, J. (1999) ‘Modernity and the politics of identity’, in
M. O’Brien, S. Penna and C. Hay (eds) Theorising Modernity: Reflexivity,
Environment and Identity in Giddens’ Social Theory. London: Longman.
Sevenhuijsen, S. (1998) Citizenship and the Ethics of Care: Feminist Considerations on
Justice, Morality and Politics. London: Routledge.
Silverman, D. (2000) Doing Qualitative Research: A Practical Handbook. London:


Smart, C. and Neale, B. (1999) Family Fragments? Cambridge: Polity.
Solberg, A. (1996) ‘The challenge in child research: from “being” to “doing”’, in
J. Brannen and M. O’Brien (eds) Children in Families. London: Falmer.


Steph Lawler

Our own most cherished conceits, stubborn evasions or persistent illusions are all fashioned by
a growing stock of cultural narratives, as we try to make sense of the past and its connections
to our lives in the present. This … is what we need to study, not seek to evade. (Lynne Segal,
Why Feminism?, p. 118)

Within social research, one compelling reason for carrying out qualitative
interviews is that they offer a means of exploring the ways in which social
actors interpret the world, and their place within it. These interpretations are
often extremely complex and nuanced, and would be difficult to access
through other means. This chapter is about one way in which such interpre-
tation can be conceptualized. It deals with the ‘narrative’ dimensions of
people’s accounts within qualitative research – the ways in which people
make and use stories to interpret the world. I want to emphasize, however,
that I am not using ‘narrative’ here to indicate a ‘story’ that simply ‘carries’
a set of ‘facts’. Rather, I see narratives as social products produced by people
within the context of specific social, historical and cultural locations. They
are related to the experience that people have of their lives, but they are not
transparent carriers of that experience. Rather, they are interpretive devices,
through which people represent themselves, both to themselves and to
others. Further, narratives do not originate with the individual: rather, they
circulate culturally to provide a repertoire (though not an infinite one) from
which people can produce their own stories.
In other words, my argument here is that, not only do people often pro-
duce ‘storied’ accounts of themselves and their relation to the social world
(within and outside of the research setting), but also the social world is itself
‘storied’. That is, stories circulate culturally, providing a means of making
sense of that world, and also providing the materials with which people con-
struct personal narratives as a means of constructing personal identities. I
argue within this chapter that narratives – which I define briefly here
(though I return to the question of definition later) as accounts which con-
tain transformation (change over time), some kind of ‘action’ and characters,
all of which are brought together within an overall ‘plot’ – are a central
means with which people connect together past and present, self and other.
They do so within the context of cultural narratives which delimit what can


be said, what stories can be told, what will count as meaningful, and what
will seem to be nonsensical.
Although narratives are most closely associated with life history research,
I want to suggest here that attention to narrative can be an extremely useful
way of conceptualizing the kinds of accounts people produce in qualitative
interviews more generally. Narratives, whether personal or ‘public’, neither
begin nor end in the research setting: they are part of the fabric of the social
world. Nevertheless, the research setting is one arena within which narra-
tives can be elicited and explored. If research participants are given sufficient
space within research based on interviews, it is likely that they will produce
narratives if asked about their lives (Riessman 1993). These narratives may
be fragmentary and partial – indeed, it is likely that they will be – but can
nevertheless tell us a great deal about the person and the social world she or
he inhabits. As always, the kind of knowledge we produce through research
will depend on what we set out to find. So, research which explores the nar-
ratives people produce will necessarily be interpretivist in nature: it will
work from the basic premise that individuals and groups interpret the social
world and their place within it. The question is less ‘What happened?’ than
‘What is the significance of this event?’ (White 1996).
From this perspective, conceptualizing interview accounts in terms of
narrative – what Somers and Gibson (1994) call ‘conceptual narrativity’ – can
be seen as a means of confounding the false dichotomy by which an inter-
viewee’s account is conceptualized either as an unproblematic reflection of
lived experience or as ‘a distorting screen that always projects experience out
of its own categories’ (Ezzy 2001: 24). It is not that ‘the facts do not matter’:
nor is it the case that ‘only the facts matter’.1 Rather, facts (or experience) and
the interpretation of those facts (or that experience) are envisaged as neces-
sarily entwined.


Conventionally, the study of narrative has been associated with literary
texts, in which context the study has largely centred on the technical com-
ponents of the narratives themselves. However, recent social-scientific work
has drawn attention to the significance of narratives for a study of the social
world. In this respect, this recent work is drawing on a relatively long tradi-
tion in social science, from the call for a ‘sociology of biography’ made by the
nineteenth-century philosopher Wilhelm Dilthey (1976), through the ethno-
graphic work of the Chicago School (which elicited individual biographical
accounts) and life history work to contemporary qualitative research which
draws on fragments of narrative (see e.g. Simonds 1992, 1996; Rickard 2001).
From the start, such work, although disparate in many ways, has been in
marked contradistinction to positivism, which would hold that the only
social facts worth consideration are those phenomena which are directly
observable and (in some sense) measurable.2 From the perspective of narrative

or voting (Ricoeur 1991a). Somers and Gibson (1994). Consider.. as I think. original emphasis) Somers and Gibson point up some important issues in relation to the study of the social world. As Lois McNay argues. but it has managed to exist within it. and not others. for example. Further. that stories guide action. that people make sense of what has happened and is happening to them by attempting to assemble or in some way to integrate these happenings within one or more narra- tives. a position which postulates that social life is itself storied and that narrative is an ontological condition of social life . public and cultural narratives. and that people are guided to act in certain ways. on the basis of the projections. but work to produce (what we count as) social reality. both connected with the ‘self-identity project’ of the social sciences: first. if narratives are not merely . Without a consider- ation of this interpretivism. This is. suggest that researchers and theorists adopt what they call ‘conceptual nar- rativity’. Second. greeting a friend. This same act could be understood in a variety of ways. variously mean. for example. It is this anti-positivism which leads Somers and Gibson (1994) to charac- terize narrative as the ‘epistemological other’ of sociology. expectations. because social scientists have tended to limit their definition of ‘narrative’ to that of merely a representation.244 CHOICES IN CONTEXT inquiry. They suggest two principal reasons for this disavowal. The study of narratives may have been peripheral in sociological research. that ‘experience’ is constituted through narratives. (Somers and Gibson 1994: 38–39. ‘Meaning is not inherent to action but is the product of interpretative strategies amongst which narrative is central’ (McNay 2000: 95). they rather overstate their case that sociology is premised on not being narrative. an action as simple as that of raising an arm. the issues with which studies of narrative are typically concerned – issues of ontology and identity – have been defined as outside of the remit of sociological concern. in brief. following Paul Ricoeur. what is apparent may be not simply inadequate but misleading. These developments include. that people construct identities (however multiple and changing) by locating themselves or being located within a repertoire of emplotted stories. what is apparent – what can be seen and observed – is not all there is to say: a much more interesting issue is that of interpretation – how social actors interpret the social world. Somers and Gibson (1994) argue. They argue. that recent developments within sociological work have made the study of narrative both more pressing and more central. Against this position. in other words. even if. narratives have conventionally been treated merely as (more or less trans- parent) ‘carriers’ of a set of ‘facts’: they have not in themselves been seen as social products. however. that the entire project of sociological theorizing and research has been premised on a disavowal of narrative. and their place within it. It is fair to say.. In other words. an increasing atten- tion to ‘textuality’: to the ways in which written (or spoken) texts are not simply unproblematic ‘reflections’ of some pre-existing reality. and memories derived from a multiplicity but ulti- mately linked repertoire of available social. for example. depending on the context: it can. hailing a taxi.

In this context. I am following the work of Paul Ricoeur (and later writers who have followed his conceptualization of narrative) who. asocial beings. 3. Plots are not selected a priori. nor as entirely determined by social structures. for example. Ricoeur defines emplotment. self and other (Stanley and Morgan 1993). I will argue that this identity is not isolated from the social world: rather. individual and society. an archive. original emphasis) Although Ricoeur’s definition of emplotment may seem pitched at a rather high level of abstraction. the central element of a narrative is its plot. on the other hand. the synthesis between two different senses of time. though. He elaborates three main (and overlapping) forms of synthesis at work in emplotment: 1. selves are acknowledged as neither entirely unique. the synthesis between many events and one story. as open and indefinite. ‘time is both what passes and flows away and. but are produced through the process of emplotment. Also relevant here is an increasing attention to issues of selfhood. ‘and then? and then?’) and time as something closed and ‘over with’. To recapitulate. Time. It is emplotment which turns disparate events into ‘episodes’ (Somers and Gibson 1994) which have . to a chronicle. con- tributing to the progress of the narrative. But these components must be brought together within an overall plot. the plot both incorporates quite disparate events and episodes. at its most basic. I address these issues: first. 2. the synthesis between discordance and concordance. and functions as a totality: one story. and so on. This leads to a radi- cal questioning of the conventional binarisms of sociology: structure and agency. Building on his argument. plot line and char- acters. as ‘a synthesis of hetero- geneous elements’ (1991a: 21). Increasingly. what endures and remains’. an event is more than simply something that happens: it must have a part in the story. it is intimately bound up with the social world. including an overturning of conventional modernist assumptions about the self. In other words. subjectivity and identity. It must contribute to the coherence and intelligibility of the narrative. unintended consequences. more than anyone else. In later sections. embedded in a series of successive incidents (so that we ask of a story. For Ricoeur. the main point is relatively straightforward: that it is emplotment which makes an account a narrative. has consid- ered narrative as a category by which people make an identity. narrative must contain transformation. on the one hand. As Ricoeur puts it. rather than simply taken as given. (Ricoeur 1991a: 22. EMPLOTMENT I have already given a brief definition of narrative – of what makes an account a narrative (as opposed. NARRATIVE IN SOCIAL RESEARCH 245 carriers of something else (something more important) then they must be subject to scrutiny. or a set of questions and answers) and here I want to expand on this definition. I consider how narratives work through processes of emplotment.

As a mode of explanation.246 CHOICES IN CONTEXT a part in the beginning. but also through space and through social categories) and I do not intend it to stand for all narratives. Conversely. and for connecting those two apparently ever-divorced categories – ‘the individual’ and ‘the social’. .3 and is taken from a much longer dialogue throughout which Barbara narrates her class movement. though. And it is emplotment which translates events into episodes. a narrative must have a point: as both Paul Ricoeur (1980) and Carolyn Steedman (1986) have pointed out. are not narratives. Rather. questions and answers and so on. Within narratives. causal emplotment is an account- ing (however fantastic or implicit) of why a narrative has the story line it does. but it must be meaningful. as significant. nor do they hinge on an overall ‘plot’. rooted in personal and social histories. inevitably. The end of a story does not have to be predictable. (Somers and Gibson 1994: 58) The significance of events is. Significance is con- ferred on earlier events by what comes later. socially. in which time passing structures the narrative and earlier events are seen to cause later ones. narratives become naturalized as the episodes which make up the ‘plot’ appear inevitable. whether the sequence of episodes is presented or experienced in anything resem- bling chronological or categorical order. as they consider the usefulness of narrative inquiry for connecting the past and present of the social actor. In short. and the end of the story is understood as the culmination and actualization of prior events. and even universal. Before this. I want to try to make the issues more concrete by introducing a frag- ment of a narrative. it is the absence of emplotment which means that archives. and in social ‘intelligibility norms’ (Gergen and Gergen 1986) – in what gets to count. It is part of a set of three interviews with one woman. NARRATIVITY: A CASE STUDY The narrative I discuss here is taken from my research with mothers of daughters (Lawler 2000a). self-actualization and movement. I want to emphasize that this is a speci- fic type of narrative (a narrative about self-development. and through processes of emplotment. the question every narrator tries to fend off is. they must incorporate this important element of bringing together disparate elements into a single plot: The connectivity of parts is precisely why narrativity turns ‘events’ into episodes. ‘So what?’ And for narratives to have a point. Barbara. These forms do not incorporate the important element of time. not only through time. my inten- tion is to use this extract as an illustrative device. prior events seem inevitably to lead to later ones. the end and the movement of a plot. Later sections will explore these issues. In this sense. Even if the events seem unrelated they will be brought together through the overall coherence of the plot.

The activities I have. .L. that’s what I am. and considers herself to have now become middle class. I don’t know – [long pause] S.. I identify with that. Action and characters are minimal. the way I decorate it. I do. is more interesting. well. As a little girl. Does that sound dreadful? S. for me at least. I’d hate to . as Barbara’s identity (and her ‘lifestyle’) is changed – even though. as we will see. why? Like eating fish and chips on the prom sort of thing. All that sort of thing I did away with immediately [when she got married]. . in this case. You know. I always wanted – er. I always had the one that wasn’t broken when we were bombed out [in the Second World War]. centres on a transformation through class movement. I always wanted to get out. to go back. knee high. NARRATIVE IN SOCIAL RESEARCH 247 Barbara grew up in a working-class family. .: No. my pastimes. The plot incorporates Ricoeur’s concept of synthesis. . her job.. Yeah.4 Barbara: You know. her house. yes! [laughs]. well he accepted it. and I feel – hmmm. . Alternatively. And I love it. you know. I enjoy being . actually. And a range of disparate elements is brought together within an overall plot – one which. It’s not back – there’s no back and forward. this idea of ketchup on the table. And my husband. And it was always Barbara’s. but nevertheless present. Barbara: Oh good. . . largely through marriage to a middle-class man. .: Tablecloths? [laughs] Barbara: Tablecloths. . you know. Barbara related many expressions of pain and inadequacy at lacking a history of being middle class. But that was me. making it one characterized by untruth. but now he’s beginning to think. are all – I suppose could be classed as middle class.. I didn’t want anything to do with that and he’s got a habit of doing that and [her daughter] Julia’s got a habit of doing that. Her account generally unsettles many conventional notions of ‘class iden- tity’: although she considers herself middle class. ’Cause I just feel dreadful sometimes saying it. This is part of Barbara’s account of her classed movement. – I’ve changed so much. from. and Barbara always had to have the same spoon. everybody else had these plastic eggcups. It would be impossible to read off Barbara’s own (narrativized) identity simply from the ‘external’ markers conventionally used – for example.5 Yet what. for example.L.. though very brief. middle class. (added emphasis) There are several ways in which this extract could be interpreted: Barbara’s words could. My lifestyle’s so dif- ferent. The extract. . the sort of house I have. the change is to something she con- siders herself to already have been. her income. There were fine things in life that I wanted to appre- ciate [laughs].. be seen as ‘innocent’ purveyors of some prior truth. . It was china. is there? To not have the . has the components of a narrative: it has transformation (change over time). is the issue of interpretation – the ways in which Barbara uses (culturally circulating) stories to interpret her life. her post-hoc version of events could be seen as ‘biasing’ her account. and it was decorated.

that is. But Ricoeur’s insight is to argue that the logic which relates these elements to each other is not that of nature. Events come to seem ‘logically’ related to each other through time. They must mesh with the rest of the world and with other people’s stories. but of narrative itself. THE SIGNIFICANCE OF NARRATIVE: LINKING PAST AND PRESENT Memory is not like a video record. the story we tell of our life is reshaped around them. the change in how meals were served when Barbara got married. but their real role is the cre- ation of a life. one total- ity (the story of what Barbara is like and of how she came to be how she is). It does this within a synthesis of two different aspects of time. some attention must be given to the epistemological status of such stories: what kind of knowledge do they represent? What kind of knowledge is being produced when conceptual narrativity is used? Memory is notoriously unreliable. 1995) has called ‘memero-politics’ – a process by which the past is interpreted in light of the knowledge and understanding of the sub- ject’s ‘present’. in asking people to recount (whole or partial) stories.. a character. at least in externals. Carolyn Steedman (1996) points out the obvious. for the object is altered and changed by the very search for it. the only point lies in interpretation. but little-noted. We must constantly engage in recall. For example. and images are never enough: moreover. by weaving stories about our past. inter- pretation. by what we call memories. It does not need images. … We constitute our souls by making up our lives.248 CHOICES IN CONTEXT as it synthesizes many disparate events (eating fish and chips on the prom. circum- stances and characteristics as rooted in childhood events. The tales we tell of ourselves and to ourselves are not a matter of recording what we have done and how we have felt. so that Barbara’s early life. or of individual psychology. our memories shade and patch and combine and delete. But the point doesn’t lie there. retelling. back in the lost time at which they happened. . although the story is not. which really cannot be found.6 Most Euroamerican read- ers will read linear progression into narratives (especially life narratives). her child- hood attachment to an eggcup and a spoon) into one overall story. the best analogy to remembering is storytelling. but the issue here is one of more than faulty versus accurate memory. (Hacking 1995: 250–251) Any research which aims for more than a snapshot of the social world must be able to somehow account for the relationship between the present and the past. The . for example. ‘[t]he search is for a lost object. not least because Euroamericans tend to conceptualize adult traits. back in the past. Time seems to move forward through the story. narrated in a linear way..7 As Steedman puts it: We all return to memories and dreams … again and again. fact that the past is no longer here: in searching for the past. is both closed and ‘over with’ and enduring into the present. in fact. into something quite different and strange’ (Steedman 1996: 103). in order to conjure up the past: we must engage in what Ian Hacking (1994.

it also makes ourselves intelligible’ (Moore 1994: 119). assimilating a life into a story is not so natural or so obvious as it might appear. identity is something which is produced through narrative. Henrietta Moore argues. or narratives. either for the researcher. social actors constitute a life. Every time we write a CV. Identity is not a pre-given entity. For Ricoeur. Partial life stories will capture some part of that interval. are a means by which people make sense of. of our relationship to that world. (Stanley 1992: 61) So too with other forms of research: neither researcher nor researched can fully access or inhabit a past which is inevitably gone. Conventionally. this process should be subject to scrutiny. on to which . or give an account of ourselves to a friend or an acquaintance. nor to a straight- forward and unmediated ‘experience’. or indeed for the research subject. it is through such stories that we produce identities. my argument here is that attention to narrative provides a way of con- ceptualizing links between past and present (rather than seeing such links as unproblematic and straightforward). both to ourselves and to others. develops and changes. So what is the relationship between living and narrating? Between a life and a life story? Through process of emplotment. and in the process. So. Liz Stanley (1992) points up this gap between past and present in her analysis of auto/biography: she argues: the ‘self who writes’ has no more direct and unproblematic access to the ‘self who was’ than does the reader. and live their lives. (Steedman 1986: 5) There is no unmediated access to the ‘facts of the matter’. as Ricoeur notes. Further. A complete life story characterizes the interval between birth and death. and that interpretation of it can only be made with what people know of a social world and their place within it. Indeed. Ricoeur (1991b) calls this ‘narrative identity’ and Somers and Gibson (1994). and anyway ‘the autobiographical past’ is actually peopled by a succession of selves as the writer grows. it is interpreted and reinter- preted in the light of the person’s knowledge and understanding. As the past is remembered. or explain inherited characteristics. and it is through such stories that we make sense of the world. the problem of ‘stories’ is seen to centre on ‘bias’: is the subject misremembering. Both are referring to processes by which people construct identities that are relatively coherent and stable through using stories. understand. constitute an identity. But. … If narrative makes the world intel- ligible. we are engaged in story-telling. Stories. and producing an identity which is relatively stable across time. and of the relationship between ourselves and other selves. ‘ontological nar- rativity’. ‘narrative is a strategy for placing us within a historically constituted world. the past is constantly worked and reworked to provide a coherent sense of the subject’s identity. The argument being made here is that we all tell stories about our lives. NARRATIVE IN SOCIAL RESEARCH 249 past is re-used through the agency of social information. misrepresenting or simply lying about earlier events? But this accent misses the point: there is no ‘unbiased’ access to the past.

But what looks like a natural.250 CHOICES IN CONTEXT narrative structures are (more or less violently) imposed. In Ricoeur’s formulation. and interpreting the self in terms of the emplotment of events: Memory . (Ricoeur 1980: 186. it works to naturalize the plot. from what could be to what is. a post-hoc treatment. .. We ‘read time backwards’. according to Ricoeur. the episodes which make up the plot are there because of the purpose they serve (which is. from present to future. among them. original emphasis) Narrative. Although this is. itself – that apparently most natu- ral and asocial of phenomena – is profoundly social. through anecdotes and episodes. as Ricoeur puts it. the very constitution of an identity is configured over time and through narrative. both connotes and constitutes movement – the movement from the potential to the actual.. brings us back to the almost motionless constellation of potentialities that the narrative retrieves. from past to present. The self is understood as unfolding through episodes which both express and constitute that self. He writes: the self does not know itself immediately but only indirectly by the detour of the cultural signs of all sorts which are articulated on the symbolic mediations which always already articulate action and. The hero is who he [sic] was. interpreting later events in the light of earlier ones. Emplotment configures a self which appears as the inevitable outcome and actualization of the episodes which constitute a life. The end of the story is what equates the present with the past. In the process. inevitably. then. making later events seem the natural and inevitable culmination of earlier ones. a teleological one: that is. but something produced through the narratives people use to explain and understand their lives. Gayatri Spivak puts it neatly: ‘We “write” a running biography with life-language rather than word-language in order to “be”. Narrative mediation underlines this remarkable characteristic of self-knowledge – that it is self-interpretation. (Ricoeur 1991b: 198) Identity. causal relationship is. ‘read- ing the end into the beginning and the beginning into the end’ (Ricoeur 1980: 183). producing a coherent plot): ‘Looking back from the conclusion to the episodes leading up to it. What we make of ‘experience’ depends on what we know about the ways in which those experiences relate to the wider social circum- stances of our lives. is not something foundational and essential. and is always-already interpreted. then. Call this identity!’ (Spivak 1997: 359). the narratives of everyday life. Rather.8 In narrating a story. What the plot configures is how the person comes to be who he or she is (as in Barbara’s narrative). we have to be able to say that this ending required these sorts of events and this chain of actions’ (Ricoeur 1980: 174). the actual with the potential. the self. is itself the spiral movement that. social actors organize events into ‘episodes’ which make up the plot. then. the events of a life come to appear as though naturally and inevitably leading to their specific conclusion.

The stories produced by individual social actors would make no sense if they did not accord. ‘the end is read into the beginning and the beginning into the end’ (Ricoeur 1980: 183) as Barbara narrates an identity that is stable across time. above. interpreting the past through the lens of social informa- tion. the rise of disciplinary frameworks (Somers and Gibson 1994) and stories produced within and about legal frameworks . If we return to the extract from Barbara’s account. because narratives of individual lives must always incorporate other life nar- ratives: hence the myth of the ‘atomized individual’ is immediately exploded. in this case. coherent and meaningful overall ‘story’. it is clear that Barbara is producing this narrative. earlier events gain significance only through events that come later. They might also be seen as ‘traditions’ (Taylor 1989). Narrative links together the individual and collective in two ways: first. and using this information to formulate present and future life stories. ‘that’s what I am’). we can perhaps rework Kierkegaard’s famous pronouncement that ‘Life is lived forward but understood backwards’ (quoted in Geertz 1995: 166). this (interpreted) past informs the present. narratives are not only produced by individuals. and. Through narrative Barbara is able to constitute a more or less coherent self:9 one which. then equally. people are not free to fabricate narratives ‘at will’ (Somers and Gibson 1994. as the connectivity of personal narratives comes to the fore (Stanley 1992. the movement of the plot leads to her becoming what she is (‘that was me’. These narratives include stories produced to explain scientific ‘dis- covery’ (Myers 1990). Ewick and Silbey 1995. Despite expressions of change. as we envisage life as something both lived and understood for- ward and backward in a ‘spiral movement’ of constant interpretation and reinterpretation. not from the past. is marked as always-already middle class. NARRATIVE IN SOCIAL RESEARCH 251 In this sense. People constantly produce and reproduce life stories on the basis of memories. further. with broader social narratives. Emplotment brings together past and present events in (what is understood to be) a logical. Indeed. Miller 1999). but out of the present. She is interpreting her earlier life in the light of what she now knows. although they do not have to be particularly ancient. to constitute an identity. I have argued that people use narratives to interpret the social world and their place within it. however obliquely. Narrative provides a means of conceptualizing people in the context of history: if the past is always interpreted through the present. Second. However. Somers and Gibson (1994) refer to such narratives as ‘public narratives’ – narratives which are ‘attached to cultural and institutional formations rather than the single individual’ (Somers and Gibson 1994: 62). McNay 2000). THE SIGNIFICANCE OF NARRATIVE: LINKING THE INDIVIDUAL AND THE COLLECTIVE So far. but also circulate socially. Hence. and as I have indicated.

One particularly compelling public narrative for contemporary Euroamericans is that of childhood ‘development’. but because the for- mer explanation accords with what Gergen and Gergen (1986) call the ‘intel- ligibility norms’ of this particular culture. narratives of development. Hence ‘we’ can be drawn in to the story. foreclosing certain kinds of story). Euroamericans almost inevitably look to childhood as the grounds of adulthood. and. The point. becoming . rather than the authoritative voice of the expert (Franklin 1990). The story she constructs is one about class and. for example. original emphasis). this perspective tends to underplay the workings of power in the social world: it is important to stress that public narratives are powerful in structuring the kinds of things which can be said (and. for this reason. define what constitutes an appropriate or successful narra- tive’ (Ewick and Silbey 1995: 207. as they operate within different cultural and institu- tional settings. for example. For most contemporary Euroamericans. in some ways. equally. expert knowledges. both must rely on certain assumptions about what is to count as ‘true’ and ‘reasonable’. any (personal or public) story which used genetic explanations as a means of causal explana- tion would be likely to be accepted far more readily than one which relied on. using its framework as a schema of self-understanding. Barbara’s narrative illustrates this link between personal and public nar- ratives. in which later (adult) psychologies are inevitably rooted in earlier (childhood) events. She is able to do this because these items are not only ‘things’ but also saturated with meaning. It is not. but this link is obscured through narrative’s use of the inclusive voice of the story-teller. that the narrative of childhood development is ‘true’ while that of heterosexual romance is ‘false’. It is scarcely possible to formulate a coherent life story without using. say. in many cases. spoons. However. Both types of narrative can incorporate different kinds of interpreta- tion. Her story is coherent precisely because it resonates with broader public narratives and symbolic systems. Barbara uses her desire for ‘fine things’ to mark out the distinctiveness of her childhood self. but. She takes several mundane items of everyday life – tablecloths. conversely. and also to mark out that self’s trajectory from childhood to her present age. more specifically. rather.252 CHOICES IN CONTEXT (Ewick and Silbey 1995). eggcups. ketchup on the table – and con- structs a story around them. Narrative is frequently linked with author- itative. is that they become ‘truths’ through their frequent repetition across a range of sites and. class movement and class desires. Paul Ricoeur argues that the kinds of public narratives on which people draw to frame their personal narratives are largely those of the literary tradition. childhood is ascribed a ‘special’ status. the narrative of heterosexual romance and marriage remains compelling even in the face of rising divorce. The point about public narratives is not their facticity or otherwise. They become symbols within her story. Similarly. Ewick and Silbey (1995) write: ‘Consent rules. through their association with ‘expert’ disciplines (such as psycho- logy). witchcraft as a means of structuring events – not necessarily because the hearer (or reader) has first-hand evidence either way.

2000a. Skeggs 1997. Further. Her individual story is narrated within a collective set of meanings.g. Skeggs 1997). NARRATIVE IN SOCIAL RESEARCH 253 meaningful precisely through their part in a narrative. In other words. But it is important to ask what kinds of power are working through rendering some kinds of narrative ‘noble’ and others ‘trivial’ or ‘petty’ (or indeed. 1993. If people interpret the past ‘through the agency of social information’. This is espe- cially so for women whose ‘upward’ class mobility has been effected through marriage to middle-class men. as Steedman (1986) argues. then that social informa- tion is going to be. no readily available female equivalent of ‘the working-class boy made good’ narratives popular- ized in the post-war period (Steedman 1986. is located within a classed system within which both her working-class background and her desire to become more middle class are frequently marked as pathological (Walkerdine 1990. Walkerdine 1997. from which Barbara wants to distance herself. This. they must set in place the conditions in which people are likely to produce narratives. Lawler 1999. are not?10 USING NARRATIVITY The task of researchers using conceptual narrativity is multilayered: first. similarly. contingent on their social posi- tioning. There is. but differential social positioning is reflected in the kinds of stories people tell. Social location is significant here.) ‘Ketchup on the table’ and ‘eating fish and chips on the prom’ signify much more than types of food and where they are eaten. Both she and I are aware that we can easily be labelled pretentious. there are few narratives which are available to Barbara. I think. any kind of narrative around class is difficult for her and always risks the opprobrium of being labelled ‘pretentious’ (Bourdieu 1984). She is able to use these symbols because both she and I inhabit a culture in which they are saturated with classed meanings. 1999. 1997. Barbara. Bourdieu 1984. Not only is there a limited range of stories which can be told. ‘What people know of the world’ is not free floating: it is a product of their specific social location. and clearly straightforward observation will not do it either. Some narratives will not easily ‘fit’: not all will afford an easy recognition. Second. for example. they circulate culturally to signify class. pretentious). for example. the symbolic meaning of these things does not originate with Barbara: rather. (But see e. Whose lives are given worth through these processes? Whose lives. . is why Barbara (very reasonably) looks for reassurance at the end of the extract above. to narrate her classed movement. I shall not go into the cultural configurations of class here: the point is that Barbara uses widely circulating cultural symbols to produce her own narrative. So Steedman (1986) presents her working- class childhood as one which did not readily accord with hegemonic narra- tives of ‘normal’ (middle-class) family life. to some degree at least. as a woman. They signify a specific kind of classed location. Reay 1998. A structured questionnaire. 2000b. conversely. Walkerdine 1997). Hence. is unlikely to do this.

Chapter 10 in this volume). will to some extent influence the kinds of narrative research subjects produce in the research itself. Further. they have to consider the relationship between these public narratives and the personal narratives produced. they do speak certain ‘truths’ about people’s (socially located) lives and identities. this does not mean. although fairly minimal in this extract. the researcher must consider the processes of interpretation going on. It is important to reiterate here my earlier suggestion that narratives are not produced. to some extent. and which operate as constraints on the kinds of narrative they can produce. then they are also being reworked in the research itself.254 CHOICES IN CONTEXT researchers have to analyse those accounts in terms of narrative. does. or as ideological ‘traps’. stories are actively worked up by social actors. move the narrative along. the issue here is not one of ‘bias’ or ‘distortion’. though not. from an endless repertoire of social narratives: only some things can be said at all. CONCLUSION I have argued in this chapter that conceptual narrativity offers a way of con- sidering the kinds of stories which people tell without considering such stories as either expressions of an ‘authentic’ identity.11 Again. for example. Thus. Rather. ab initio. that narratives are simply ‘imported’ more or less wholesale into the research setting. people fre- quently combine fairly contradictory forms of narrative in producing their personal narratives. The ‘truths’ people produce through such stories are not ‘truths’ as conventionally understood in positivist social science: nevertheless. in the research setting. the whole direction of the research. In Barbara’s narrative. from this perspective ‘identity’ is not something which can be read off . Narratives do not have to be lengthy or full accounts of a life: they simply have to incor- porate the processes of emplotment outlined above. of course. of course. Above all. Rather. Through such narratives (and within social and institutional constraints) people produce identities (however fragmented. they are co-produced between the researcher and the research subject (see Jennifer Mason. people may postulate a self which is the product of genetic inheritance at the same time as postulating a self which is the product of social shaping. if narratives are constantly worked and reworked. but one of the inevitability of interpretation and reinterpretation. But research. So. In this sense. However. Third. This relationship need not be harmonious or smooth: the process of constructing a personal narrative is creative work (even if it is not self-consciously so) and people may well use public narratives only to oppose them. they are drawing on competing narra- tives. The kinds of questions asked. And fourth. introduces another level of interpretation – that of the researcher. both personally and publicly. yet combining them in a new narrative form. my own intervention. either. multiple and contingent). In this. researchers need to consider the kinds of publicly circulating narratives on which social actors draw. as will the location of the researchers themselves. and only some things can be said from any specific social and histori- cal location.

unified and coherent self may be a fiction. 5 The dichotomy to which I referred in the Introduction at the beginning of the chapter. it seems to me. are as linear as we might assume. ‘Any story. she talks of her classed identity as ‘two aspects of your life. of course. however truthful it aims to be. 2 Although. This is why I characterize the self which is constituted here as more or less coher- ent. ACKNOWLEDGEMENTS My thanks to Steve Fuller and to Paul Johnson for comments on earlier drafts of this chapter. or as independent of events. Although the identities people hold are certainly related to the social context they inhabit. categories. see Ezzy (2001). relatively unproblematically. At the same time. people’s interpreta- tion of the world cannot be assumed from these categories. six dots (……) indicate a pause. If we want to find out how people make identities. NARRATIVE IN SOCIAL RESEARCH 255 from an externally imposed schema. I doubt that many positivists would accept a narrative interpretation of events. the process of being any- thing is. Paul Ricoeur. . despite his hermeneu- tical stance ‘is entirely willing to accept the products of positivism’ (Erben 1993: 17). 8 Although ‘word language’ is also necessary. she suggests. they may not). however. 7 See also J-B. of course. more complicated than simply an identification with single. in order to represent this running biography to ourselves and to others. Barbara’s own problematizing of a linear model of time (‘It’s not back – there’s no back and forward. make sense of the world and of their place within it – if we want to find out how they interpret the world and themselves – we will have to attend to the stories they tell.12 What is more. is a reconstruction from the vantage-point of the present’ (Pontalis 1993: xv). 9 The whole. people may well be able to accommodate levels of non-coherence within their self-narratives. externally imposed. A full account of the methodology can be found in Lawler (2000a). ‘women’ and so on (although. People may well belong. My aim here is to consider the process as much more entangled than either of these polarities allow. but it is one with which many people – perhaps most of us – work. to groups designated ‘working class’. NOTES 1 Narratives have been approached in many different ways in social research – often precisely as either simply ‘reflecting’ events. which you try to juggle’ (Lawler 2000a: 109). 4 In the extract that follows. but this in itself does not tell us about the kinds of identities they build. 3 ‘Barbara’ is a pseudonym. For a good overview. And if this is the case. the narrative Barbara presents unsettles the idea of a coherent self: elsewhere. then the doing of qualitative research becomes pressing indeed. In many ways. as Michael Erben (1993) rightly notes. is there?’) highlights this point – neither lives nor stories. 6 Indeed. who argues. Pontalis.

Gergen. Doing. Bourdieu. I. Four Decades. Bourdieu. 12 For example. Ezzy.256 CHOICES IN CONTEXT 10 Although this is not the place to explore the subject in detail. Durham. for exam- ple argues that notions of resistance tend to approve only behaviour approved by the bourgeois observer – behaviour which ‘not only resist[s] domination but [does] so for decidedly progressive aims’ (Fox 1994: 8). Ewick.M. C. London: Macmillan. In my own work (see Lawler 2000a) I found that the mothers I interviewed often disidentified from the category ‘mother’. M. Writing. Are they condemned to being characterized as ‘reactionary’ (as. S. . Franklin. and only did so after a long discussion of classed position and classed movement. W.J. Dilthey. (1999) The Weight of the World: Social Suffering in Contemporary Society. in T. D. translated by R. (1995) After the Fact: Two Countries. the ‘respectability’ espoused by many working-class women (Skeggs 1997) may be held in contempt precisely by those who have the cultural capital to shrug off respectability. it is worth noting that ‘resistance narratives’ may themselves be extremely restrictive. K. in M. Sociology 27(1): 15–25. Cambridge: Cambridge University Press. St Leonards. and Silbey. New York: Praeger. Cambridge: Polity. (1993) ‘The problem of other lives: social perspectives on written biography’. the working-class women in Beverley Skeggs’ (1997) study frequently disidentified from the category ‘working class’. My own located- ness is. London: Routledge and Kegan Paul. working to approve some types of action and diminish others. One Anthropologist. (2001) Qualitative Analysis: Thinking. NSW: Allen and Unwin. McNeil. indeed. REFERENCES Bourdieu. Fox. (1976) Selected Writings. and similar feelings of occupying a class ‘hinterland’. at least in a more optimistic time. I think. (1994) Class Fictions: Shame and Resistance in the British Working-Class Novel 1890–1945. MA: Harvard University Press. P. 11 Barbara was embarrassed to tell me some of these things. S. (1990) ‘Deconstructing “desperateness”: the social construction of infertility in popular representations of New Reproductive Technologies’. Sarbin (ed. Geertz. working- class women often have been)? Ironically. NC: Duke University Press. P. M. Cambridge: Polity. (1995) ‘Subversive stories and hegemonic tales: toward a sociology of narrative’. Law and Society Review 29(2): 197–226. Erben. P.R. But it must be asked what happens to those whose narratives cannot easily fit this template. P. Barbara and I share a similar class background. P. (1993) The Field of Cultural Production: Essays on Art and Literature. edited and introduced by R. the demise of capitalism). (1984) Distinction. Nice. important here. Johnson.) Narrative Psychology: The Storied Nature of Human Conduct. Cambridge. Pamela Fox (1994). and Gergen. Yearley (eds) The New Reproductive Technologies. (1986) ‘Narrative form and the construction of psychological science’. This may be why narra- tives of (for example) ‘the heroic male worker’ can seem so attractive (signalling. Varloe and S.

London: Routledge. translated by D. seen it. in L.R. (1998) Class Work: Mothers’ Involvement in their Children’s Primary Schooling. G. G. Daughters. Stanley. Nash (ed. W. W. Cambridge: Polity. Ricoeur. Segal. Calhoun (ed.D. NJ: Rutgers University Press.R. (1996) ‘Consuming selves: self-help literature and women’s identi- ties’. McNay. B. S. . Wood (ed. London: Sage. J-B. Ricoeur. Wood (ed. Cambridge: Polity.L. Lawler. (2001) ‘“Been there. (2000) Gender and Agency: Reconfiguring the Subject in Feminist and Social Theory. trauma and the soul’. Thousand Oaks. and Morgan. Manchester: Manchester University Press. Hacking. Riessman. P. Rooney) (1997) ‘In a word: interview’. L. NARRATIVE IN SOCIAL RESEARCH 257 Hacking. Lawler. H. Feminist Review 63: 3–24. Skeggs.) Narrative in Culture: The Uses of Storytelling in the Sciences.) Social Theory and the Politics of Identity. Reay. London: Routledge. (1999) Researching Life Stories and Family Histories. Moore. London: UCL Press. New York: Routledge. (1993) Narrative Analysis. (2000a) Mothering the Self: Mothers. (1993) Love of Beginnings. (1992) The Auto/biographical I. Psychology. K. D. M. (1986) Landscape for a Good Woman: A Story of Two Lives. (1980) ‘Narrative and time’. London: Routledge. and Gibson. Spivak. Cambridge. Ricoeur. Critical Inquiry 7(1): 169–190. C.) Subject to Change: Cultural Studies and the Working Class. Stanley. R.) On Paul Ricoeur: Narrative and Interpretation. I. P. Lindlof (eds) Constructing the Self in a Mediated World.K. G. Change and Social Worlds. Pontalis. Sociology 27(1): 1–4. London: Routledge. NJ: Princeton University Press. W. (1991a) ‘Life in quest of narrative’. Feminist Review 67: 111–132. Philosophy and Literature. S. (1999) Why Feminism? Gender. (1997) Formations of Class and Gender: Becoming Respectable. London: Routledge. Rickard. (1991b) ‘Narrative identity’. Wood. History of the Human Sciences 7(2): 29–52. D. (2000b) ‘Escape and escapism: representing working-class women’. London: Virago. hopes. London: Free Association Press. MA: Blackwell. L. Myers. in C. London: Sage.) The Second Wave: A Reader in Feminist Theory. the future and retirement’. (1995) Telling Sexual Stories: Power. Cambridge: Polity. (1999) ‘Getting out and getting away: women’s narratives of class mobility’. (1992) Women and Self-Help Culture: Reading between the Lines.S. (with E. in C. London: Sage. CA: Sage. (1994) A Passion for Difference: Essays in Anthropology and Gender. Subjects. (1993) ‘Editorial introduction’. P. Plummer. C. (1995) Rewriting the Soul: Multiple Personality and the Sciences of Memory. I. (1990) ‘Making a discovery: narratives of split genes’. Munt (ed. in S. in D. Somers. Steedman. in D. L. Politics.) On Paul Ricoeur: Narrative and Interpretation. (1994) ‘Reclaiming the epistemological “Other”: narrative and the social constitution of identity’. L. I’ve got the t-shirt”: British sex workers reflect on jobs. Simonds. S. London: Cassell. Grodin and T. New Brunswick. Miller. Nicholson (ed. in D. Lawler. (1994) ‘Memero-politics. done it. Princeton. Simonds.

(1989) Sources of the Self. Walkerdine. (1996) ‘Commentary’. MA: Harvard University Press. (1990) Schoolgirl Fictions. White. C. History of the Human Sciences 9(4): 123–138. London: Macmillan. H. (1996) ‘About ends: on the ways in which the end is different from an ending’. Walkerdine.258 CHOICES IN CONTEXT Steedman. V. . History of the Human Sciences 9(4): 99–114. C. V. (1997) Daddy’s Girl: Young Girls and Popular Culture. Cambridge. London: Verso. Taylor.



processes. Voyle and Simmons 1999. social scientists are working with community groups and policy-makers to undertake and support the evaluation of a range of community-based projects (Connell et al. There is often a presumption among many researchers that choosing qualitative methods will enhance the potential for participants’ own voices and values to inform the research process and thus illuminate in a grounded fashion the evaluation of a project or policy. that is the focus of this chapter. of qualitative evaluation (Chelimsky 1997. The overall aim of this chapter is to assess the potential for qualitative research to address critically issues around power and participa- tion in evaluation. Certainly a qualitative approach can focus on accounts of the ori- gins and progress of events. In this way it is argued that a quali- tative approach ensures a degree of participation not possible with a quanti- tative perspective. It is the relationship between qualitative methodology and evaluation. and the implications of this for the actual conduct of an evalu- ation. happiness and well-being’ and the project is located in five of the . Crucial to these expectations is competing perspectives on power and the potential for an evaluation to truly engage with communities and other stakeholders. if not the actual conduct. Yet these claims can raise expectations among research participants about the conduct of an evaluation that cannot or may not be met. and thus the complexity and multidimensionality of the local contexts and relationships. In many of these projects there has been a consideration. Underpinning this presumption is a view that qualitative research can provide a more informative picture of culturally based processes and practices and a depth to context-based expla- nations of events. Sonas is the Gaelic word for ‘good fortune. To illustrate and illuminate debates on qualitative methodology and evaluation I draw upon my experiences of and data from two qualitative evaluations: • SONAS Community Health Project. outcomes and ultimately future policy and practice. These developments are part of a more general trend to evaluate complex social interventions and community work as a means of informing policy and practice. Shaw 1999).12 ENGAGEMENT AND EVALUATION IN QUALITATIVE INQUIRY Linda McKie Increasingly. Wallerstein 1999). 1995.

1995). local authorities. This Action on Smoking and Health (ASH) Scotland project was funded by HEBS and operated between 1996 and 1999. voluntary sector organi- zations. Subsequently. The chapter opens with a brief discussion of definitions of evaluation and qualitative evaluation. data from the two evaluations are presented to inform a critical discussion of qualitative methodology and evaluation. • Women. the provision of one-to-one counselling work. which also funded an evaluation of the first three years’ work. Nineteen community-based initiatives across Scotland were funded to explore new ways of working on smoking reduction among women living on low income. In the former I was commissioned by the funding body to coordinate the evaluation and in the latter project I was a member of the evaluation advi- sory group. Both of these projects are evaluations of community health projects. and revisits some of the ‘basic set[s] of beliefs that guides action’ (Guba 1990: 17). 2001). More than ever before community groups. Low Income and Smoking Project. consi- ders how those needs can be met and decides collectively on priorities for action (Connell et al. Communities are complex entities and are comprised of heterogeneous groups and stakeholders. and the UK National Health Service are being . RESEARCH AND QUALITATIVE EVALUATION At a fundamental level the evaluation of social interventions is concerned with making an assessment. EVALUATION. the type of community health work undertaken dictates the evaluation approach and methods. and workshops and conferences (SONAS 1996). In the final section the ideas and practice of qualitative evalua- tion are considered with specific reference to the relationship between the evaluator and stakeholders and recipients of the programme. Invariably. of judging the merit of an activity or pro- gramme and assessing it against the goals that were established at the out- set (Barlow et al. There were nineteen community-based projects funded in two waves to explore new ways of reducing smoking among women on low incomes. Community health work describes processes by which a group in a given locality (or organized around a specific topic) defines its health needs. In a broader sense evaluation seeks to address practical problems and make judgements of merit or worth so as to provide recommendations and outcomes that may inform future activities (Connor 1993). PARTICIPATION AND EXPERTISE islands in the Western Isles of Scotland.262 POWER. The project commenced in 1993 and until 1996 was funded by the Health Education Board for Scotland (HEBS). Illumination of the processes by which the various constituencies in commu- nities negotiate and refine agendas on health and how these are implemented in related activities tends to lead to the adoption of qualitative methods. During this period the project provided a range of health activities rang- ing from group work.

outcome-orientated measures of accountability. evaluators and project participants recognized and struggled with issues of power. Some commentators and researchers suggest that qualitative evaluation should be used only when quasi-experimental quantitative designs will not work (Chelimsky 1997). Thus these are useful examples to explore the ways that commissioners. But for the proponents of qualitative evalua- tion it is the ‘particular challenges and constraints facing evaluation [that] usu- ally require qualitative evaluation as a methodology of choice’ (Shaw 1999: 5). and to ensure accountability. subtlety. These assessments were thought to provide the opportunity to verify what worked. in both of the projects discussed in this chapter evaluation was built in from the beginning and active participation of the various groups in setting both the research and practice agendas was promoted. However. valid and reliable easy-to-follow results and implications for future work. or otherwise. particularity. Qualitative methodology affords a means of providing distinct data and qual- itative evaluation of theorizing problems and approaches. educational and community policies. (Shaw 1999: 1) While a qualitative approach has become increasingly popular in the evalua- tion of a range of activities. Many are uneasy with the ambiguities posed by qualitative data especially in the context of the sub- jective relationship between evaluator and those participating in the project and hence evaluation research. Shaw (1999) comments: quantitative evaluations. It enables stake- holders and project recipients to highlight and reflect upon what worked and how this came about. are typi- cally bereft of the intensity. ethical judgement and relevance that potentially characterize evaluation born of ethnography. Yet many of these evaluations have been criticized for failing to bring to life not only the everyday workings of policies or projects but also the lessons that might be gained from reflecting upon processes. for example. Nevertheless it would be naive to presume that the . Quantitative methods are often used in the belief that these could provide practitioners and professionals with clear-cut. many evaluations commence well into the implementation of a project or policy and thus leave limited time and space for the stakeholders and recipients to take part in the design or conduct of the research and evaluation. Nevertheless. for example evaluations of penal. Controlled evaluations of projects or policies were often assessed against behavioural and other measurable outcome criteria. ENGAGEMENT IN QUALITATIVE RESEARCH 263 required to build in evaluation from the outset of projects or policies in order to gain reliable information about activities as well as evaluating processes and outcomes (Wimbush and Watson 2000). not least of which is the evaluation of community- based projects and policies. many policy-makers and practitioners continue to favour quantitative approaches. Contemporary evaluation research has its origins in the push to develop evidence-based action in educational and social policies during the 1960s. working with a range of groups and professionals to establish a project or implement a policy. ownership and participation. and affords an opportunity to chart and reconcile multi- ple stories of a project.

Drawing upon Miles and Huberman (1994) and Shaw (1999) there are a number of opportunities and characteristics to consider: • contact with participants is possible at various levels in a manner likely to provide a depth. divergence and minority argu- ments become evident. Quali- tative research can ‘provide important insights into different perceptions of “reality” although debates continue as to whether the researcher should view “reality” from the outside or appreciate their role in the construction of social meaning’ (Guba 1990: 27). political and economic structures (Bryman 2001: 278). dialogue and delibera- tion in all stages of planning. nominal group processes such as the Delphi technique (Grbich 1999: 116) may actually diminish a sense of participation or ownership of the research process. there is a general agreement that evaluation makes judgements on the merits of specific policies or programmes (although there is a lot of debate on approaches to evaluation) while research is concerned to theorize and describe issues or problems. This process is repeated until areas of consensus. • it is possible to develop a relationship between data at various levels from the ‘on the ground’ participants to policy-makers and funding bodies (Shaw 1999: 90). data collection and analysis and consideration of outcomes and recommendations. • analysis and writing up is represented in words and based upon judge- ment. Not all qualitative methods promote participation to the same degree. Thus conclusions can be open to challenge as subjective (Shaw 1999: 179). and to pro- vide a spatio-temporal context to these (Bryman 2001: 277). • qualitative approaches ‘can effectively give voice to the normally silenced and can poignantly illuminate what is typically masked’ (Greene 1994: 541). • there are opportunities to gain a holistic overview by locating accounts in a context of a social intervention and wider social. Qualitative methodology offers opportunities for evaluation research that it would be useful to outline.264 POWER. These concerns further raise the need for the application of qualitative methods in a fashion that reflects upon and seeks active inclusion. There are differences between the two that emerge with reflection upon the purposes of each (Shaw 1999: 11). For exam- ple. Debates abound on the epistemological underpinnings of qualitative research that range from an emphasis on methods and techniques to paradigmatic perspectives (Grbich 1999). Crucial to qualitative evaluation research is an appreciation of the distinct nature of evaluation and of research. However. As a first stage this technique involves a panel’s identification of a range of issues on a topic and from that a series of questionnaires are sent out to panel members to further consider their opinions on issues raised. to accounts of events. PARTICIPATION AND EXPERTISE adoption of qualitative methodology necessarily enhances participation. Most significantly. as well as breadth. the Delphi technique has been criticized for creating a notion of ‘experts’ and often involving those who developed a pro- ject or policy in a process that aims to consider and influence change. .

As Tracey Reynolds (Chapter 14) demonstrates. Denzin and Lincoln (1994) argue that only since the mid-1970s have notions of objectivity and value-free research been challenged by post- modern and cultural studies (Lincoln 1994). ENGAGEMENT IN QUALITATIVE RESEARCH 265 Qualitative evaluation is appropriate where potential outcomes are not clear in advance. and of great relevance to evaluation. Lofland and Lofland (1995) note the ethical dimensions to qualitative research. but also it has been criticized as reflecting the views of the researcher/evaluator who may be driven by idealism or realism borne out of the aesthetics of pervasive accounts of groups or individuals. where commissioners are interested in appreciating and understand- ing social processes and processes of participation in the project. Within contemporary debates researchers can propose a reconceptualization of evaluation research and produce research designs and processes that challenge the maldistribution of resources and power between researcher and researched. in the policy arena its value continues to be questioned and muted. Crucial to the design and con- duct of qualitative and participatory evaluation is transparency and negoti- ation. there must be agreement among the commissioners of the evaluation. often in comparison to quanti- tative work. practitioners and participants with the approach and methods. However. Finally. Time and resources must be available to explore a range of data sources and not just the most obvious. This can involve both qualitative and qualitative methodologies. Without a consideration of the role of power differentials in establishing agendas for a project and an evaluation. not least of which. not every researcher nor commissioning body is comfortable with ongoing review and renegotiation of data sources. data analysis and the evaluation process. power in social research is ‘multifaceted. Add to this the critical assessment element of evaluation and the relationship between the people who are giving their time to provide data and the evaluator of that time becomes imbued with the differential operation of power. qualitative evaluation can appear transparent but will merely reinforce inequities in decision-making and pro- ject activities. and diverse data sources are available. where a project may involve numerous groups or issues. for example. POTENTIAL AND PITFALLS FOR A PARTICIPATORY EVALUATION Qualitative research may involve a greater commitment from the researched in that data collection frequently requires an infinite and in-depth relation- ship to be developed between themselves and the researcher. is the ques- tion ‘is it ethical to take sides or to avoid taking sides in a factionalized . while qualitative evaluation may be perceived as producing material that can be useful and potentially powerful in giving voice to the concerns and ideas of a range of people and groups. Not only is its value questioned. relational and interactional and is constantly shifting and renegotiating itself between the researcher and the research participants according to different contexts and their differing structural locations’.

I consider the dilemmas posed when the evaluator and project team sought to actively tackle notions of ‘power over’ through supporting initiatives to undertake evaluation themselves. there are instrumental goals of evaluation that pose particular ethical and practical problems. Participation in qualitative evaluation is critically explored in the action research evaluation undertaken of the SONAS Community Health Project. Shaw (1999: 140) notes that participation in an evaluation requires a clar- ity and mutual agreement on a number of matters such as the purpose of the participation. As Wallerstein (1999: 42) argues. dance together. Huberman (1991) proposes three types of relation- ships that can provide barriers to participation and collaboration: • ‘Hello–goodbye’ where researchers and communities meet. Wallerstein 1999). There are myriad interpretations and workings of any community-based project. ‘with increasing participatory evaluations. Raising expectations of evaluation. a recognition of the varying interests of participants. makes it difficult to develop a clear purpose and evi- dent participation in evaluation while avoiding fears that qualitative evalu- ation is an opaque way of making a judgement on local input and involvement. In the second example.266 POWER. then wander off. the breadth of groups and coalitions mean that there is no one single project to evaluate. Concerns about participation in evaluation raise specific issues for quali- tative evaluations. input and responses to data analysis. mechan- isms for exchange and communication. values and expectations. Low Income and Smoking Project. On this last point participation can place a further burden on people whose perceptions of evaluation may be that of a judgement on their work or lives. particularly the potential to participate in dynamic contexts. ongoing feedback on the final report and dissemination (Shaw 1999. and can further complicate the effort of general input to a project. PARTICIPATION AND EXPERTISE situation?’ While authors have sought to address ethical issues in qualitative evaluation it is essential for the evaluator to state clearly where they stand on issues such as participation in the design and conduct of the project. community members have raised ethical challenges. wanting evaluations to be useable and accountable to the community. Huberman (1991) has suggested that there is a ‘two community’ problem with the researcher and the researched community having differing norms. or summative infor- mation that is received too late to be useful’. token feedback. Changing social. the ownership of data. A major issue is the multiplicity of definitions of ‘community’ and the voices within spatial or interest based definitions of community (Peterson and Lupton 1996). the evaluation of the Women. . economic and political environments also present additional challenges to the evaluator as these will impact on the project and shift the content and context of community health work. While community-based qualitative research and evalua- tion may seek to generate information and knowledge. instead of abstract reports. and an appreciation of the benefits or potential drawbacks of involving people in the evaluation.

Lynne Haney (Chapter 13) calls for feminist scholars to think more seriously about ‘the politics of reflexivity in different research settings’. Further he criticizes what he perceives as an emphasis in qualitative research as ‘empowerment under the aegis of subtle benign oppression’ and ‘luring gate- keepers and informants into being studied by a design in which they are not invited to collaborate. Acknowledging our position as social scientists and our potentially unique broad perspective on social life. and to which. There are those in projects (stakeholders and recipients) who may know about certain kinds of research and thus gain a power base in negotiations of agendas. those for whom evalua- tion is a relatively unfamiliar activity may recognize the potential to draw a critical assessment from the research and consider resisting or posing barriers to further involvement in the project or the evaluation. In prac- tical terms the negotiation of agendas to establish a qualitative evaluation may be characterized by opposing and conflicting ideas on methods and approach. they are only invited to give informed consent’ (Heron 1996: 28). amongst other contexts. at best. • ‘Stand-off’ where there is little agreement and some resentment between communities and researchers. As researchers we may locate ourselves in terms of gender. These are major challenges to the power/knowledge and relationships between the researcher and . Haney argues for a more contextual approach to reflexivity that considers. Postmodern research proposes a dialogue and praxis between researchers and partici- pants to question issues of power and reciprocity (Wallerstein 1999). ENGAGEMENT IN QUALITATIVE RESEARCH 267 • ‘Two planets’ where neither group has much contact with each other. and geography and locality. Communities may seek to establish their power and knowledge bases and struggle with evaluators and commissioning bodies to tackle those who appear to exert ‘power-over’ them and strive for notions of ‘power-with’ (Wallerstein 1999). class and status in relation to others. race. community life and historical legacies. Therefore establishing and reviewing the aims of and format for an evaluation is an ongoing and necessary process for mutual appreciation of the possible benefits of qualitative evaluation. commissioners and participants (factors we often try to dis- miss from academic debate on research and evaluation). institutional structures. By contrast. The type of relationship that will develop in any evaluation is dependent upon a range of factors including the abilities. capabilities and character of the evaluator. Yet we cannot anticipate how those involved in the research and evaluation may perceive the role. authority and ultimately the power of the evaluator. controlling research on people and fully participatory research with people’. Heron (1996: 27) asserts that cooperative inquiry may not achieve participation and that ‘qualitative research about people is a half-way house between exclusive. resources. aims and objectives of the evaluation. In her exploration of two ethnographic studies she outlines how she opted to conduct her research not from an inward perspec- tive of autobiography but instead to turn outward establishing and assessing the impact of the wider social world on the participants and issues more gen- erally.

and the report of the case is likely to have a theme. Projects were established in communities undergoing a series of economic and social shifts and with cultural and religious histories. These contexts and shifts necessitated an appreciation of tensions within commu- nities. Low Income and Smoking Project. issues arose which were of relevance to the various localities and the respective projects.’ Marshall and Rossman (1989) go on to detail the purpose of the case study being to chart. on the basis of this. between agencies and issues of communication and power between evaluators. EXPLORING PARTICIPATION AND POWER IN QUALITATIVE EVALUATION The findings and outcomes of the evaluations are detailed elsewhere (ASH Scotland and HEBS 1999. PARTICIPATION AND EXPERTISE evaluator. While both evaluations were of projects premised upon a community development approach to health work. Case studies ‘examine the bounded system of a program. reflexivity and dialectic between researcher and researched are starting points in revealing the dynamics of power and participation in qualitative evaluation. and reflecting on this material across different levels of participation in the project. These aims sound worthy but are not without debate and controversy as discussed below. advocating power analysis. so recipients and stake- holders might use the evaluation to inform future work. In the evaluations discussed below action research and case study approaches were adopted. These issues are first explored through experiences and data derived from SONAS: Community Health Work in the Western Isles (McKie 2000).268 POWER. . has a unique life. However. It is an approach that emphasizes collaboration and partici- pation. depict and characterize events and activities. In this account of an evaluation I highlight the classic dilemma for the qualitative evaluator. The case. facilitators and participants. In the second evaluation discussed in this chapter. in some ways. the decision to adopt this approach and its implications for qualitative evaluation are considered. Action research is a means of gathering data about a project. believing I might promote the potential to develop ‘power-with’ communities. McKie 2000) and brief descriptions of the projects and evaluations are presented here. the Women. Stake (1995: 133) proposes that ‘the case to be studied probably has problems and relationships. For Atkinson and Delamont (1993) the issue of what counts as a ‘case’ is potentially problematic. but the case is an entity. In my role as coordinator for the evaluation I created concerns among the funding body and manage- ment groups about the objectivity of the evaluation research and content of the final report. an institution. to adapt or change the project on an ongoing basis (Hart and Bond 1995). These concerns arose from my attempts to address notions of ‘power-over’ the community in the evaluation. or a population’ (Marshall and Rossman 1989: 44) with the pur- pose being ‘to reveal the properties of the class to which the instance being studied belongs’ (Guba and Lincoln 1981: 371). The overall aim of this approach is to generate knowledge and.

where possible. The SONAS Community Health Project commenced in 1993 and was funded by HEBS for three years to April 1996 (£40. Roads. The extra daylight of the summer months is maximized for family and home-based activities. ENGAGEMENT IN QUALITATIVE RESEARCH 269 SONAS Community Health Project The islands on which the project is located cover an area of approximately 70 square miles and the distance between South Uist and Stornoway (the administrative centre for the Western Isles) is 120 miles. The remoteness of the islands is exaggerated in the win- ter months with less frequent transport services and poor weather. a sizeable minority have two or more jobs which vary according to the seasons. educa- tion. In addition. there are tensions between so-called ‘incomers’ and locals. These tensions reflect economic and cultural differences. fish farming or building. This pattern of life provides mechanisms through which the loneliness and boredom of the winter may be addressed. and work in tourism and related services. Being an incomer suggests a person or family who have settled in the islands to take up employment (often professional and better paid jobs) or to retire. These islands are situated to the west of mainland Scotland and are the last landmass before Newfoundland. It is costly to travel between the islands and within the islands bus services are extremely limited. tourism and fishing – or derived from public sector bodies – health services. based on the concept of community health volunteers • devise and set up. Also many men and women work on a seasonal and part-time basis. specifically the Western Isles communities • develop and pilot innovative methods of health education and promo- tion. Employment is either related to the landscape and sea – crofting. mechanisms and structures to facili- tate the involvement of local communities in assessing and tackling health education priorities • make recommendations for future health and social policies • develop resource materials and guidelines for use by other agencies or organizations wishing to learn from the project and devise similar schemes • evaluate and disseminate the work of the project. causeways and ferry and air services interconnect the islands.000 per annum was allo- cated for each of the three years). Private sector employment is minimal and often linked to sole or joint trading in crafts (predominately jewellery manu- facture). subsequently the Western Isles Health Board took over funding. The overall aims of the project were to • identify the health education needs of populations in remote and rural areas. Yet the long winter months are characterized by many social and cultural activities. Unemployment is evident across all the islands. As a result a car is a necessity to access most local services. HEBS funded the project under its ‘demonstration’ project scheme to explore the potential to learn from the community development approach to health . regional council and army.

Objectivity and power: perceptions of and dilemmas for the evaluator The SONAS project and evaluation were set up in the first three months of 1993 as the funding body. By this time the evaluation had been put in place and initial research work undertaken by Lisa Curtice and Andrew MacGregor. HEBS. semi-structured interviews and questionnaires comprising open-ended questions. PARTICIPATION AND EXPERTISE work as applied in rural areas. as well as creating further problems. A major aim in these early years was to gain as much involvement from various dimensions of the island communi- ties as possible. As coordinator of the evaluation for the project I took up my role in late 1993. putting the abstract notion of community development principles into con- crete action as a health project began to mirror tensions within the islands. The evaluation was a formative and summative one. In addition. Participants in the project were encouraged to take an active involvement in the monitoring. wished to ensure that monies were . As a result the process and outcomes of the project were examined on an ongo- ing basis and on the analysis of data the evaluation team undertook regular feedback. the project became active in local economic and planning debates on the sustainability of these remote and rural communities. A project coordinator and a part-time administrator were appointed.270 POWER. Western Isles Health Board and the Scottish Community Development Centre. coro- nary heart disease and stroke and mental health. conferences. In 1996 the Western Isles Health Board took over funding for the project and this decision was based upon the outcomes of the evaluation. A range of access points were developed from one-to-one counselling. However. and sharing this material with the evaluation team (McKie 2000). making videos of activities and taking photographs. Small area census data were also utilized to provide a socio-economic profile of island life (McKie 1996). to chronic illness and mental health. In addition. Whenever possible the project coordinator continues to develop work with a community development ethos. reflection and resultant change processes. project participants were encouraged to record activities. The project spent the first two years of funding developing a community agenda for health work and existing groups and activities combined with newly formed groups to consider a range of health issues from men’s health. The overall aim of the evaluation was to provide a critical assessment of the work. workshops. inter- viewing each other. Drafts of data analysis were returned to many respondents and data regularly discussed with the project advisory and partners groups. manage- ment structure and organization of the SONAS project. employing an action research approach. open meetings and group work. The project con- tinues but has become more focused upon major health themes of cancer. Qualitative methods were used: focus groups. A partners group was set up to act as an advisory and steering group and this comprised a representative from HEBS.

The coordinator of the first year of the evaluation did make contact with many local groups and people but most were hesitant about involvement early on. These views were further complicated by an admission on the part of HEBS that SONAS was a mechanism for delivering health promotion in remote and . While local activists expressed an early interest in the project. All the key players in funding. In addition. The community hospital in one of the islands was being considered for clo- sure and this reinforced a sense of outsiders failing to consider local agen- das and the potential role of local community decision-making. Nor was there time to consider the local cultural and socio-economic contexts. ENGAGEMENT IN QUALITATIVE RESEARCH 271 allocated by the end of a given financial year. local people were unsure what the project was about and initially concerned about becoming too involved. Many appreciated the ethos of the approach but had witnessed these projects come and go. prior to the launch of the project there was no time for consultation with local people on either the content of the project or methodology of the evaluation. In subsequent research for the evaluation local people commented that the launch meeting was ‘very much run by out- siders’ (identified as the partners’ representatives) and did not provide clear guidance on developing project work. initially there was little contact between the com- missioners. funders. given the geographi- cal location of the project. For example. with local people unsure what was expected of them at either the meeting or in the pro- ject and the evaluation. Many spoke of staff from the administrative centre in Stornoway ‘jetting in and out’. The project was launched in the summer of 1993 at a public meeting held at a local hotel. instead preferring to ‘wait and see what takes off’. In the 1980s a number of community development projects had been located in the islands on such topics as economic regeneration and childcare. monitoring and evaluating the project were based in a range of organizations in mainland Scotland or in the island’s administrative centre 120 miles away. the meeting proved to be tense. the evaluation team and local people. • Local debates on the role of the Western Isles Health Board. The role of these volunteers was unclear and many local people described this idea as ‘cheap labour’ and a possible ‘cost-cutting’ approach by the health board. These views were further complicated by a sense among many local people that the project designers and the eval- uation team had failed to appreciate: • Previous experiences of community development projects. project coordinator (newly in post). The speed with which the project was set up had repercussions for communication and for dialogues between groups that impacted on much of the first two years of the project and the evaluation. the SONAS project was to recruit community health volunteers. Invitations were sent to local organizations and the event was well attended. So while assuming knowledge of the approach among local people the partners had failed to consider how previous projects had actually worked. once funding ceased there was little left behind despite the rhetoric of ownership by the community. However. In addition.

Even at this stage changes had taken place in key members of local groups. held inappropriate notions of a rural idyll with close-knit communities and that these views had led to inadequate preparation for the project and evaluation. the speed with which the project had been set up had not allowed for local consultation or reflection on the social and economic contexts. Communities are heterogeneous and tensions evident between those who have come to the islands to take up employment (often better paid professional jobs) and local families who have had to adapt to changing economic opportunities of the islands. • Diversity in and between islands. While these views and changes were recorded and fed back to the partners’ group and local people they illuminated the reality that the evaluator represented and mediated several power bases. The qualitative approach to collecting and recording information on the cultural and local contexts was emphasized. Certainly the evaluation was shifting in response to changes in the project and personnel. local people questioned the com- munity development ethos of the partners and the evaluation. • Economic concerns. While the partners’ group agreed that the project should be led by a com- munity agenda and ultimately ‘owned’ by local people. . There are religious and cultural differ- ences between the islands in the north of the Western Isles. Some expressed con- cerns that ‘outsiders’. as was trans- parency over data analysis and the process of regular feedback. The islands in the north are predominately Protestant in religious and cultural make-up. With rising unemployment rates and the contraction of an army base on one of the islands there were major concerns about the economic fortunes of the islands. including the partners’ group. Yet ironi- cally. There was a sense. As a consequence some questioned the need for a health project as opposed to economic regeneration work. the partners’ group and in the evaluation team. that the project and evaluation were being imposed on communities and that this was unlikely to promote participation in either the project or the research.272 POWER. strongly vocalized by local people. Issues raised about the role of the evaluator Much of the first year of the evaluation was spent winning local people over to the approach and methods – convincing them that their voice could have an equal weight with others. some felt that the project would prioritize health board interests rather than those of local people. A local health survey was suggested and local work on this was conducted but funded from other sources. In many ways the community development ethos provided a strength of approach to the project. at the earliest stages of the project. while the islands in the south are largely Catholic in religious outlook and atti- tudes. PARTICIPATION AND EXPERTISE sparsely populated localities. and the administrative services are located on one of these islands. Yet it was difficult to deal with the community’s direct criticisms of the early stages of the project and evaluation.

We just don’t want to run the risk of losing anything that brings something to these islands. While many appreciated I was not working for the funding body. ‘Was that OK? What I said won’t lose anyone their job will it?’ Similar comments were addressed to me after two focus groups conducted with local health groups. I emphasized my own position as a researcher. Being Irish. Many local people perceived the evaluation team and the process of the evaluation as another means of collecting information from ‘insiders’ to judge the project and to promote decisions by ‘outsiders’ on the future of the work and the extension of the three-year post of project coordinator. as the evaluator I had the authority of the partners’ group. If I wanted people to speak to me then I would have to adapt to this cycle of local activities and so I changed the schedule for the research and the evalu- ation. My pattern of working was at odds with the seasonal nature of local life. However. I also had the authority of being an academic working from a university on the mainland. For the duration of the evaluation some participants were extremely cautious in their interaction with me. I came into the island communities for a week at a time but went home to a somewhat more secure economic situation than most of those from whom I was collecting information. especially the funding body. and having moved a number of times to study and to work in islands proud of their Celtic . employed at a university and commissioned as an independent evaluator. HEBS. One local woman said this ‘was noth- ing personal. I was never fully able to overcome the sense that I was part of a critical and poten- tially negative decision-making process rather than an evaluator informing the future of the project. Initially I used acade- mic terms to describe the evaluation and this appeared to illuminate notions of evaluation as a critical assessment. This was brought home to me after completing an interview with an active participant in the project who commented. Early on I recognized that island life was organized around the weather in a way that challenged my assumptions that the best time for me to collect data would be the spring and summer months. While this made some aspects of the evaluation work and processes more accessible. ENGAGEMENT IN QUALITATIVE RESEARCH 273 First.’ After this comment I began to feel a sense of responsibility for the future of the project and the need to make sure any decision to change or end the project was firmly grounded in research evi- dence. I had demonstrated a lack of knowledge and awareness of local culture and patterns of activities. One means of addressing these concerns was to ensure transparency in the process of data analysis and involve as many participants as possible as well as members of the partners’ group. Yet I did have several factors in my favour. I was told in no uncertain terms that little would happen with the project over the summer but rather that a whole raft of cultural and social activities would start in the autumn and run over the winter. not everyone shared perceptions of my role as independent and separate. This prompted me to talk instead of recording a history of the project and of local responses to it. Involvement in these various groups and meetings gave people a reason to socialize and tackle the potential isolation of the long winter months.

PARTICIPATION AND EXPERTISE traditions and from which many migrated. Low Income and Smoking Project aimed to explore new ways of working with a community development approach to address smoking reduction among women living on low income (Gaunt-Richardson et al. Distributing a total of £37. participants and the commissioning groups. A variety of methods were used by the initiatives to provide an oppor- tunity for participants to address their smoking behaviour while develop- ing new skills and interests and raising their confidence and self-esteem. I was also working at a university located in the north of Scotland. 1999). I consider the dilemmas posed by actively seeking to tackle notions of ‘power-over’ and propose ways to develop ‘power-with’ community-based initiatives in evaluation. it was difficult to overcome the sense that the evaluation was imposed. focusing on cessation. Though some initiatives tackled smoking using direct methods. Nevertheless. a member of the partners’ group was concerned with the number of interviews and focus groups conducted with local people. Also being female and a mother gave other avenues to relate to some of the project participants. Women.000. in that negotiation I demonstrated a lack of knowledge on local culture and concerns. Approaches to evaluation by the respective initiatives also demonstrated a diversity of methods and resources (Amos et al. In contrast. . exercise and diversionary activities. and that while ongoing attempts were made to renegotiate the content and schedule of the research. the Women. The project also sought to gain new insights into appropriate ways of evaluation by drawing upon the skills and experiences of funded initiatives. McKie et al. the majority used a variety of indirect methods such as poetry and video work. with all the travel difficulties associated with that. as a means of engaging with women to tackle the issues around smoking. and this ensured that I visited the islands for at least a week at a time. For example. Local people tended to see me as generally supportive of the project but part of a decision-making process on the future of the pro- ject in which they had little say. 1998). 1999.274 POWER. 1999. nineteen initiatives were awarded grants of between £500 and £3. the partners expressed concern that I might be too involved in local matters and thus less objective. did provide some common ground. drama. Barlow et al. To a degree I avoided the tag of ‘jetting in and out’. the second of the qualitative evaluations. This member of the group regularly queried the health knowledge and skills of local people and their ability to comment upon the content of the project and thus suggested that too many data were collected from sources less well informed than others. Funded by HEBS and run by ASH Scotland. I was never able to overcome contradictory perceptions of local people and partners.000 to undertake work that would operate for a maximum of a year and provide a final report on evaluation. Low Income and Smoking Project In this.

methods and learning. It was felt that working with initiatives and drawing upon their strengths. participa- tion. and working across groups and organizations. The evaluation advisory group was established early in the life of the project to provide general support and advice both to the project manager and to the facilitators of individual initia- tives. Two informa- tion days. 1998. given the above it was further presumed that the evaluation advi- sory group would compile the evaluation of the overall project as final reports were submitted. The final report was to be as short or as long as initiatives wished but did have to provide a full report of activities and the views of participants. sustainability. Four project bulletins were produced to disseminate information about the funded initiatives to groups and organizations throughout Scotland and a final report circulated nationally and inter- nationally (Gaunt-Richardson et al. that most community groups would have previously acquired skills in evaluation. ASH Scotland and HEBS 1999). ENGAGEMENT IN QUALITATIVE RESEARCH 275 Those funded agreed to record their work and produce a final report that would also form an evaluation of the initiative. a seminar day and an end of project conference were organized to bring initiatives together at different stages of the small grant funding process so as to enhance opportunities for initiatives to share their work plans. It was hoped that reports might also consider such issues as planning. Second. first. The skills of the advisory group were to be accessi- ble to initiatives at the information days and accessed at other times through . A project manager and administrator were appointed. Although 20 per cent of the grant award was withheld until submission of the final report. a management group established to oversee the project work and an evaluation advisory group set up to support the evaluation work of the initiatives. Its role was not to undertake evaluation research per se but to work towards extracting overall conclusions from the final reports of the initiatives. This group was appointed after the project format and content had been agreed and a number of assumptions about evaluation had already been made by HEBS and ASH. A database detailing information from all the small grant applications was set up to provide general information and data for the evaluation. while recogniz- ing that limited support may be required. namely. especially given the increasing trend to accountability in the voluntary and public sectors. would be an approach sympa- thetic to the community development ethos of the project. Further it was anticipated that initiatives would come to a critical self-assessment of their work. Low Income and Smoking Project adopted a commu- nity development ethos it was anticipated by ASH and HEBS that the evalu- ations undertaken by the initiatives would be pluralistic in nature and probably utilize qualitative methods. where cash flow was found to be a problem it was agreed to review individual cases. The potential for ‘power-with’ in qualitative evaluation Given that the Women.

Thus limited resources and time were allocated in the initial project budget and timetable to support evaluation. this further reinforced the assumption that skills. the project was designed more than a year prior to its launch and in the intervening period senior personnel at ASH and HEBS had changed. resources and experiences in the community could and should be drawn upon. Even those working in an organizational context that might have staff who were trained in evaluation and who actively supported this dimension of an initiative’s work began to express concerns. The information day started with a number of presentations on the ethos of the project. and asking too much of initiatives that had day-to- day community work to develop on relatively small sums of money. An information day was held in the first months of the project and brought together the funded initiatives. At its first meeting the members of the evaluation advisory group agreed that they were keen to achieve a balance between securing meaningful infor- mation and evaluations. the processes of establishing and sustaining a group and working with other organizations left many feeling distinctly uneasy. Hence the facilitators failed to appreciate the relevance of evaluation work and actively questioned what they considered to be ‘a lot of fuss’. These arrangements were publicized in the project newsletter and by the project administrator and manager. Third. Later it transpired that this community group secured two to three grants per year from a range of funding bodies and had never submitted a final report. More worry- ingly. the variation in skills and experiences became apparent. PARTICIPATION AND EXPERTISE the project manager. One initiative questioned the relevance of the whole idea of evaluation. This success in acquiring grants was linked to the high levels of deprivation in their locality and the community group being the only one in the area. and design- ing and administrating the occasional short questionnaire were the levels of experience in most of the community groups. Recording work through the minutes of meetings. As initiatives got underway some realized that their experience in group work and community development did not necessarily equip them for evaluation. Over coffee a number of facilitators expressed concern at the time and skills they felt were required to fulfil the evaluation requirements. Therefore advice was sought from national and local bodies such as HEBS. attendance records at sessions. The idea of exploring the place of smoking in women’s lives. the project management and evaluation advi- sory groups to consider. Others asked for practical guidance on research methods.276 POWER. among other issues. the evaluation. In addition. ASH Scotland had little prior experience in evaluation and no expe- rience of community development work and the evaluation of such projects. tobacco control and evaluation. During small group discussions specifically on recording and evaluation. as the project manager visited initiatives it became evident that some . Yet in the early months of the project it became evident that to assume that past skills and experiences would equip initiatives to undertake a relatively inde- pendent evaluation of innovative community-based work was placing too much pressure upon many of the facilitators and indeed participants.

The project manager and administrator spent an increasing amount of time on sup- porting recording work and the evaluation advisory group began to meet more often. In reflecting upon these concerns. however. In conjunction with the advisory group. to support those who could do this work but might have lacked confidence. Further. ENGAGEMENT IN QUALITATIVE RESEARCH 277 were struggling with evaluation work but were trying to hide this for fear of calling into question the release of the final 20 per cent of the grant. and to ensure a level of quality in final reports that would allow overall conclusions to be drawn. namely. One initiative asked the project manager to look through twelve large files of material and notes on interviews. As a result the evaluation advisory group proposed a change to the evalu- ation approach. By early 1997 a new problem was emerging. a number of the initiatives found these extremely useful in providing a format for decisions on information to include and dimensions of their work that required reflection and critical assessment. were aimed at health promotion or community education officers and assumed a good level of knowledge of research methods. for example. especially the analysis of qualitative data from interviews and focus groups. we explored the availability and relevance of resource materials on evaluation aimed specifically at community groups. At the same time the project as a whole was gaining publicity in Scotland and across the rest of the UK and Europe. the evaluation advisory group and project manager agreed to provide specific support for several groups. one initiative received input from a consultant researcher. the project manager designed a series of evaluation questions that initiatives could consider when compiling their final report. and advise on what could be derived from this for the final report! Others also began to ask for detailed advice on analysis and what should go into the final report. Yet the project manager and evaluation advisory group began to feel increasing responsibility to relieve the pressures felt by many facilitators. Most resources. We had reached a point where the idea of ‘power-with’ ini- tiatives had been challenged. the analysis of mate- rial. Our adherence to the notion of supporting a community development approach to evaluation and support for independent evaluations by the respective initiatives had led to a series of concerns that . Change and evaluation research In the light of these events Wave 2 applicants were asked to provide fur- ther detail on their evaluation plans and encouraged by the project man- ager to spend time planning recording and evaluation work in the early stages of activities. Nevertheless the majority of Wave 1 initiatives started recording work and used a range of methods to do so. and was receiving numerous requests for further information and for detail on the outcomes and recommendations when these became available.

For the end of project report. methods and content of the project. Data collection was undertaken in 1998 and early 1999. However. These were chosen to reflect the range of settings. it was agreed to undertake five in-depth case studies of selected initiatives. The Evaluation Journey: An Evaluation Resource Pack In a final review meeting of the Women. a member of the project management group and the project manager agreed to undertake the research for the case studies and so roles changed from that of advisers to researchers. the final report • interviews with facilitators • when possible. The evaluation advisory group now had to negotiate this shift in over- all approach to the evaluation while not diminishing the potential impact of the final reports of the initiatives. So after two lengthy meetings. group interviews with participants • a review of presentations made at the information days and final project conference. The case study methods included: • a review of relevant documentation from the application form. other records. PARTICIPATION AND EXPERTISE might have been avoided with some early work on mapping the skills and experiences of facilitators. On seeking agreement from five ini- tiatives to undertake a case study it became evident that some of these ini- tiatives had begun to find evaluation so time-consuming or problematic that the idea of a case study undertaken by someone experienced in research was received as welcome support. We were careful not to diminish the evaluation work of the initiatives in the dissemination of the findings from the project. it would be difficult to draw overall conclusions on the project. It was recognized that while case studies might provide some depth to identifying and exploring issues relevant to the given ‘case initia- tive’. initiatives praised the level and quality of support received from ASH . to plan- ning documents. and if available. By this time several of the case study initiatives had completed their work. In the case of one initiative the case study formed the only evaluation as participants and facilitators just did not know how they would or could undertake this element of their work. The three members of the evaluation advisory group. Low Income and Smoking Project. Luckily the evaluation advisory group and project manager had a high level of commitment to the project and additional time available to undertake a change of role.278 POWER. key learning from a review of reports from the initia- tives was reported alongside themes emerging from the five case studies. one with the project management group. it was pos- sible to interview the majority of facilitators and in every case meet some participants.

Ultimately. suggest that a number of national and local organizations may also be overestimating the skills available. Facilitators commented that they had gained so much in the development of their skills. The epistemological underpinnings of qualitative research are often cited as promoting the potential for participation and tackling notions of power and authority between researcher and the researched. and the level of support required to undertake evaluation work at a community level were also underestimated.1 The hope by the project management and evaluation advisory groups to create qualitative evaluations that reflected notions of ‘power-with’ the funded initiatives had proved extremely difficult to realize. In addition. Their assump- tions about skills and resources were ill founded. 2001) was developed from the learning from the project and from further interviews with facilitators. The pack provides information on the rationale for evaluation. The pack is being piloted and once revisions are made publication will follow. and further. these claims must be set within . The Evaluation Journey: An Evaluation Resource Pack for Community Groups (Barlow et al. CONCLUSIONS: THE IMPLICATIONS FOR QUALITATIVE EVALUATION In the 1990s the growth in demands for accountability and the evaluation of community-based work placed a number of pressures upon researchers and communities. As illustrated in the qualitative evaluations considered above. analysis. evaluation involves making an assessment. both judging the merit of an activity or programme and assessing outcomes against potentially shifting aims and objectives. On reflection we recognized that the theory of seeking to work in an egalitarian manner may sometimes pose problems and issues for those less confident and experienced in research.2 An independent evaluator has been appointed. The evalu- ation of the ASH Scotland Tobacco and Inequality project has drawn upon this learning. ENGAGEMENT IN QUALITATIVE RESEARCH 279 Scotland. and a major recommendation to emerge from the meeting was the need for resource materials and for related ongoing general and specialist advice services for evaluation. Qualitative evaluation has been conceived of as one means of addressing the potentially problematic issues of how to make evaluation a participatory activity and thus enhance the potential for reflexivity. A major theme to emerge from the project as a whole was the need for resource materials on recording and evaluation that were both informative and accessible to community groups. in addition to the provision of materials and support from an evaluation advisory group (Barlow et al. and a number have gone on to achieve further grant awards and to undertake appropriate evaluation work. report writing and dissemination. Not every organization can provide this level of support. the analysis of qualitative data. planning and participation. collecting information using conventional and unconventional methods. 2001).

that is researchers and com- munities meet. inclusion. Identification of the groups and individuals participating in an evaluation and clarification of their respective positions on a project and evaluation must be a starting point for evaluation work. The relevance of community health work to the accountability procedures of organizations varies. be considered as ‘a halfway house’ between controlling research on people and achieving parti- cipation. Yet all too often researchers presume that by adopting qualitative methodology they have addressed the potential concerns of communities about power. Low Income and Smoking Project caused concerns among community groups who felt pressured by the demands of an independent evaluation. PARTICIPATION AND EXPERTISE the everyday context of the origins of a project. as does the appreciation of process evaluation. in attempting to address potential concerns about ‘power-over’ and ensure recognition of skills and experiences in the community the evaluation of the Women. Drawing upon the two qualitative evaluations dis- cussed in this chapter there are a number of implications I will discuss under three headings: inclusion. even if honest about the power basis to evaluation will be considered to have exercised their subjectivity on data and any final report. the epistemic basis to qualitative evaluation does afford the poten- tial to develop a dialogue and praxis. Ironically. then wander off. In its earliest days this evaluation suf- fered from a failure to undertake consultation prior to the start of data col- lection or to situate the evaluation in local socio-economic and political debates. dance together. This is perhaps the best to hope for as the researcher will almost always leave. Inclusion While qualitative research may. author- ity and participation. change and proposed content and potential outcomes from the research and evaluation (Barnes and Wistow 1992). as Heron (1996) asserts. and. Regardless of the methodology considered for the evaluation. In a review of the evaluation of the SONAS Project it was evident that an involvement in and commitment to evaluation must be established in the design stage and regularly reviewed. Despite the wish to make the evaluation as transparent as possible there was a sense of the exercise of ‘power-over’ the community among many local people. dialogue and deliberation. Most evaluators develop what Huberman (1991) has described (as we have seen) as a ‘hello–goodbye’ relationship with communities. If self-assessment is proposed the skills of those involved should be ascertained and support made available as mutually agreed. economic . a profile of communities and localities should be undertaken prior to the project or policy and evaluation commencing. and the realities facing those who commission.280 POWER. Community-based evaluation does not take place in a vacuum. Crucial to this is a consideration of understandings and values on participation. The profiling work should include detail of social. implement and participate in a project.

con- duct and analysis of the research data? • Are there plans for community input to the final report? If not. Just adopting qualitative methodology is only part of the answer to ensuring a multiplicity of voices can be heard through qualitative evaluation. These may also give rise to potentially competing interpretations of events and data and place the evaluator in the difficult situation of having to decide on a final version of events and analysis of data. among others: • Where did the original idea emerge from? • Was consultation undertaken or is the policy or project being imposed in a top-down fashion? • How much say do the community and researcher have in the design. might this be possible? • Are there mechanisms for community members or commissioners to comment upon the conduct and outcomes of the research and evaluation activities? It is all too easy to accept a commission for an evaluation. for example. Dialogue Creating a dialogue between participants. especially in the current climate that values the award of external research grants. debates about a possible hospital closure can alert evaluators to issues that may impact on the design and conduct of the evaluation. Deliberation In the analysis of data and consideration of outcomes and recommendations it is imperative to appreciate that a multiplicity of definitions of community exist and will give rise to a range of voices and opinions. Researchers must be honest about their power/knowledge bases and address the ethics of research and evaluation. espe- cially where there are external researchers or evaluators. Most importantly the process will identify groups and individuals who should be included in the research and evaluation. but the responsibilities and ethical issues can change a seemingly straightforward project into a highly problematic and stressful one. ENGAGEMENT IN QUALITATIVE RESEARCH 281 and political factors and set these in an historical context (Scottish Poverty Information Unit 1999). As Heron (1996) argues participation may not be feasible or obtainable. but there are dan- gers in asserting that participation is either not possible or that qualitative methodology will provide the appropriate context: . Profiling is likely to illuminate pertinent issues. A series of questions arise which include. funders and project facilitators requires a reflexive stance on the role of researcher and researched.

As noted earlier the evaluation of the SONAS Project assisted in the deci- sion to retain the project. as well as ensuring that . However. participation is justified by the outcomes it achieves. Although it has been adapted to reflect specific health issues the project continues to work with a community development ethos and this approach has been accepted by a majority of health board members and health care professionals. They should be ready to ensure that people derive some direct benefits from participation. Low Income and Smoking Project have led to ongoing work to promote community-based evaluation work and community development work on tobacco issues through the Inequality and Tobacco Project. particularly between the researcher and researched. 2000. The projects and evaluations have also provided further input to a number of theoretical debates concerning community development and community health work which have been discussed elsewhere (McKie 1996. as Tracey Reynolds (Chapter 14) contends. As asserted earlier in the chapter there has been presumption that qualitative research can provide a more informative picture of culturally based processes and practices in a manner that enhances the potential for a more equal relation- ship between the researcher and researched. .282 POWER. (Barnes and Wistow 1994: 91) The discussion in this chapter has used data drawn from two community health projects to demonstrate the values and pitfalls of engaging in eval- uation research. . . for example they provided an evidence base on the opportunities for and barriers to community health work through the consideration of the dynamic nature of participa- tion in community projects and evaluation. Lastly. The communities provided their own determination of ‘success’ by reflecting on the content and processes of the project and had a direct input to the research as well as the final report. The evaluations and projects themselves achieved a num- ber of successes in policy and practice terms. PARTICIPATION AND EXPERTISE [providers] should be sensitive to the demands they are placing on people whose lives are already complex and demanding. SUMMARY The overall aim of this chapter was to assess the potential for qualitative research to address critically issues around power and participation in evaluation. Recommendations from the Women. ‘power in social research is not a fixed and unitary construct’ and the realities of the shifting nature of power are evident in the evaluations discussed above. through the process of qualitative evaluation the com- missioners accepted that community involvement in the various stages of the evaluation and project was an imperative to both the development and assessment of the project. ASH Scotland and Health Education Board for Scotland 1999).

regional and local responses to the White Paper on tobacco Smoking Kills and the White Paper on health. In developing my ideas on. Paula Gaunt-Richardson. They must consider the politics of reflexivity in different research settings (Lynne Haney. Further information can be obtained from ASH Scotland. Kathryn Backett-Milburn. Chapter 13). Email: ashscotland@ashscotland. Web page: http://www. Maureen Moore. . Joy Barlow.000 to investigate new ways of working and offer new services to individuals and groups and improve and develop the way exist- ing services are offered. 1995). and work in. Scotland. • So as to illuminate local issues and structures researchers must also extend ‘outward to look more broadly at the social world’ through perspectives on ‘institutional structures. Paula Gaunt-Richardson. ACKNOWLEDGEMENTS I would like to thank all those who participated in the two evaluations consi- dered in this chapter. Emma Witney and Daniel Wybrow. Margaret Black and Emma Witney for comments on earlier versions of the chapter and to Tim May and the anonymous reader for editorial advice. 8 Frederick Street. community life and historical legacies’ (Lynne Haney. • Researchers cannot presume that the adoption of qualitative research will necessarily tackle issues such as power differentials and participation. • When evaluating or researching community-based work researchers must appreciate the multiplicity of definitions of community and power differential within communities (Connell et al. As part of this project six community-based initiatives have been funded for one year. Low Income and Smoking Project. Many thanks to Susan Gregory. and to support national. Edinburgh EH2 2HB.org. • Researchers must appreciate the role of subjectivity in the analysis and writing up and thus the need to ground assertions in various forms of evidence and reflection (Shaw 1999). namely SONAS Community Health Project and Women. to develop further approaches to community-based evaluation. NOTES 1 The Evaluation Journey. Chapter 13). An Evaluation Resource Pack for Community Groups can be obtained from ASH Scotland. ENGAGEMENT IN QUALITATIVE RESEARCH 283 Several key issues were considered in this chapter: • Evaluation is about making an assessment and this process and outcome will raise concerns and may cause sensitivities (Wallerstein 1999).uk 2 The Tobacco and Inequality Project commenced in 2000 and runs for three years with the overall aim of developing community-based services to support the reduction of smoking among those living on low income.org.uk. Towards a Healthier Scotland. qualitative evaluation I would like to acknowledge the ongoing support and advice of Kathryn Backett-Milburn. with a grant of £10.ashscotland. Norma Neill.

(eds) (1992) Researching User Involvement.. Evaluation 3(1): 97–104. (1990) ‘The alternative paradigm dialog’. Guba. A. J. Buckingham: Open University Press. Thousand Oaks. and McKie. Gaunt-Richardson. L. low income and smoking: developing community-based initiatives’. Atkinson. (1993) Monitoring and Evaluation Made Easy: A Handbook for Voluntary Organisations. Oxford: Oxford University Press. (eds) (1994) Handbook of Qualitative Research. . PARTICIPATION AND EXPERTISE REFERENCES Amos. and Lincoln. A. and Gaunt-Richardson. (1991) ‘Linkage between researchers and practitioners: a qualita- tive study’. Amos. Chelimsky. Kubisch. and Lofland. CA: Sage. (eds) (1995) New Approaches to Evaluating Community Initiatives: Concepts. C. and Moore. S. Hart. Edinburgh: ASH Scotland. Barlow. and Lincoln. Barnes. in M. Gaunt-Richardson. J. Health Education Journal 58: 329–340. E.. (1999) ‘Addressing smoking and health among women living on low income II: TAPS Tiree – a dance and drama group for rural community development’. L. (1999) Qualitative Research in Health. E. London: Sage. Connor. M. McKie. Guba (ed. Schorr. Newbury Park. Lincoln. Buckingham: Open University Press. CA: Wadsworth. Lincoln (eds) Handbook of Qualitative Research. (2001) Social Research Methods. P. (1997) ‘Thoughts for a new evaluation society’. C.284 POWER. Y. J. in N. M. (1999) ‘Addressing smoking and health among women living on low income III: Ayr Barnardo’s Homeless Service and Dundee Women’s Aid’. New York: Aspen Institute. CA: Sage. A. Gaunt-Richardson.) The Paradigm Dialog. P. and Weiss. (1994) ‘Tracks toward a postmodern politics of evaluation’. CA: Jossey-Bass. Edinburgh: HMSO. (1995) Analysing Social Settings. Lofland.. (1996) Co-operative Inquiry: Research into the Human Condition. Evaluation Practice 15(3): 299–309. Health Education Journal 57 (4): 303–313. Thousand Oaks. San Francisco. ASH Scotland and Health Education Board for Scotland (1999) Breaking Down the Barriers: Women.. A. J. L. Y. Leeds: Nuffield Institute for Health Service Studies. M. and Bond. American Educational Research Journal 27: 363–393. G. and Barlow. London: Sage. E. L. J. Bryman. A. (1993) ‘Bread and dreams or bread and circuses? A critique of case study research in evaluation’. in E. M. Guba. McKie. P. Edinburgh: ASH Scotland. and Wistow. E. Connell. (2001) The Evaluation Journey: An Evaluation Resource Pack for Community Groups.. Health Education Journal 58: 321–328. Heron.) Controversies in the Classroom.. J. Greene. (1981) Effective Evaluation. Barlow.. Denzin. Low Income and Smoking. Denzin and Y. Huberman. University of Leeds. and Delamont. (1998) ‘Women. P. P. Belmont. N. CA: Sage.. Amos. Final Report. J. Methods and Contexts. A. (1995) Action Research for Health and Social Care. (1994) ‘Qualitative program evaluation: practice and promise’. Hammersley (ed. Y. L. Grbich.

M. quality and effectiveness’. Health Education Journal 58: 311–320. CA: Sage. L.050. and Amos. Marshall. and Rossman. Wimbush. P. Community and Health: A Hebridean Conference. and Simmons. (1999) ‘Power between evaluator and community: research rela- tionships within New Mexico’s healthier communities’. Thousand Oaks. Scottish Poverty Information Unit (1999) Community Profiling Resource Pack. London: Sage. (2000) SONAS: Community Health Work in the Western Isles. Thousand Oaks. (1996) The New Public Health: Health and Self in the Age of Risk. (1995) The Art of Case Study Research. D. E. Evaluation 6(3): 301–321. Voyle. SONAS (1996) People. N. Thousand Oaks. (2000) ‘An evaluation framework for health promo- tion: theory. J. A. Social Science and Medicine 49: 1. L. Gaunt-Richardson. (1999) Qualitative Evaluation. Edinburgh: Health Education Board for Scotland.. Peterson. J. (1999) ‘Community development through partnership: promoting health in an urban indigenous community in New Zealand’. J. Glasgow: Glasgow Caledonian University. Barlow. . Social Science and Medicine 49: 39–53. and Huberman. and Watson. C. I. ENGAGEMENT IN QUALITATIVE RESEARCH 285 McKie. Miles. CA: Sage. G. London: Sage. Edinburgh: Health Education Board for Scotland. D. A. (1999) ‘Addressing smoking and health among women living on low income I: Dean’s Community Club – a mental-health project’. CA: Sage.. (1996) SONAS: An Evaluation of the Second Year of the Project. (1994) Qualitative Data Analysis: An Expanded Source Book. McKie. Shaw. R. Benbecula: SONAS Western Isles Community Health Project. Wallerstein. Stake. L.035–1. McKie. and Lupton. A. (1989) Designing Qualitative Research.

thus disrupting the androcentric bias that seemed to plague so much traditional social science. class and gender and of the ‘serial’ quality of gender. symptomatic of larger trends in feminist theory. Yet as feminists began to construct more nuanced theories of gender. feminist social scientists would enjoin them. Sandoval 1991. feminist social scientists would arrive at a clearer. The logic was clear: by beginning their inquiries from the standpoint of women.13 NEGOTIATING POWER AND EXPERTISE IN THE FIELD Lynne Haney It was in the late 1970s that feminist scholars began to critique many of the assumptions underlying social scientific research and to develop their own methodological approaches. Such arguments were clearly a sign of the times. as feminist theorists became more attentive to the intersection of race. reflexivity can imply addressing the power embedded in the researcher/ researched relationship and sharing this interpretive power with those being studied. Feminist scholars used these insights to argue for the existence of a ‘women’s standpoint’ – a perspective on the social world that was grounded in women’s shared social locations (Hartsock 1987. Instead of insisting on a separation between the ‘knower’ and the ‘known’. Our research methods would then reflect this approach. the call for reflexivity has become almost a catch phrase used to denote many different practices. ‘flexible’ standpoints. reflexivity can mean recognizing researchers’ . Much of this early work was influenced by traditional standpoint theory and its recognition of the social origins of knowledge. In particu- lar. Thus. In fact. this methodological paradigm was challenged. feminist methodology would give rise to alternative research practices. more lucid under- standing of power relations (Harding 1991). This led to the recognition that women were themselves situated differently. Smith 1987). in ways that both allied and divided them. Collins 2000). This implied the development of ‘mixed’ standpoints. As an analytical tool. they began to question the existence of a women’s standpoint (Harding 1998. Collins 2000). Whatever their speci- fic approach. most scholars agreed on the need for increased ‘reflexivity’ in feminist research. Young 1995). feminist scholars began to advance methodological approaches that took into account the similarities and dif- ferences among women (Harding 1998. and ‘partial’ per- spectives (Haraway 1991. In effect. They would transform the research experience into a dialogue and a shared project of illumination. As a research strategy.

These were cases where some women (as state actors) wielded power over other women (as clients). NEGOTIATING POWER AND EXPERTISE IN THE FIELD 287 own social locations and disentangling how they might shape the empirical analysis (Stacey 1988. I set out to conduct a series of feminist ethnographies of the state. Those attributes that some women deemed problematic were considered unproblematic by other women. power was exercised through this labelling process and subsequent determinations of what was un/problematic in everyday life. from the perspective of those caught in these systems. as Chapter 1 in this volume outlines. In my own research. I first attempted this in a study that examined the concrete interactions in a probation office and a group home for incarcerated teen mothers in California. Smith argued that sociologists would be better able to unearth ‘relations of the ruling’ and to trace ‘lineaments of oppression’ (Smith 1987: 154). I transported the model to another social arena – the state. this seemed like quite a stretch. Within these settings. In effect. In its many incarnations. In fact. it implied listening to women’s voices and transforming what they find prob- lematic into sociological inquiry. Orloff 1993. I followed these feminist methodological debates closely. Smith’s sociological approach has been used with great success in a variety of empirical studies on education and work. Pateman 1988). It also provided an alternative way to do social science: by beginning our investigations from the standpoint of the everyday. As a graduate student in the early 1990s. ‘relations of the ruling’ were enacted among women. How does a feminist scholar analyse ‘relations of the ruling’ when they occur among women? How does a . Her work addressed my sociological ‘problem with no name’ by pointing out that traditional sociology tended to take its questions from an ‘administrator’s world’. In both studies. a plethora of unexpected issues surfaced during the research process. Initially. While this methodological approach sounded promising in the abstract. feminist methodo- logy spoke to my academic concerns and desire to do social science in a different way. both femi- nist and non-feminist scholars had tended to view the state from above. This approach appeared to turn traditional sociology on its head. the goal was to construct a view of the state from below. Reinharz 1992). as a strictly macro-level entity (MacKinnon 1989. the ‘lineaments of oppression’ ran through and among women. Nevertheless. Fonow and Cook 1991. My second ethnography of the state was conducted as part of a larger project on the development of the Hungarian welfare system. Dorothy Smith’s (1987) theory of the ‘everyday world as prob- lematic’ exerted a particularly strong influence on my development as a social researcher. At the centre of these conflicting definitions of the situation were struggles over power and expertise: who knew what was best for women? What was their knowledge based on? How was it legitimized? In this chapter. This study analysed shifts in the operation of Hungarian welfare through an examination of its national-level policies and local-level practices. I discuss the practical exigencies I encountered while conducting two ethnographies of the state and how they complicated my ideas about ‘reflexive’ research practices.

and the standards of appropriate behaviour used to evaluate them. or from the policies and laws dictated from above. As I conceptualized it at the time. while the entire group home staff was female. they spoke in ‘us versus them’ terms as a way of distancing them- selves from what occurred in other justice institutions. In this chapter. Instead of formulating my questions from an administrative angle. Far more surprising were these women’s representations of their positions in the justice system. those in charge of these institutions were women: the overwhelming majority of probation officers (POs) were women of colour. in those early interviews. While the latter . Thus. I argue for the opposite approach. In and of itself. I began my first ethnography of the state in the California juve- nile justice system.2 Everyday relations of the ruling The initial phase of the research project involved interviewing those in charge of both institutions. this was not surprising. In both studies. historically. I grappled with the politics of ethnographic practice by turning outward.288 POWER. social welfare occupations have been predominately female. PARTICIPATION AND EXPERTISE feminist sociologist adjudicate among women’s conflicting views of what is problematic in everyday life? And how does a feminist researcher cope with situations in which listening to women’s voices places her on competing sides of institutional divides? Other researchers have grappled with similar dilemmas through an inward approach to reflexivity – by turning the analytical focus onto themselves and interrogating their own positionality. POs frequently opposed their work to that of the courts and juvenile hall.1 My goals were lofty: I set out to study the state in a way that took feminist methodological and theoretical principles seriously. the modes of control they were subjected to. Instead of turning inward. that the theoretical and methodological complications began to surface. it meant pulling back from the ethnographic pre- sent and moving toward ethnohistory. I grounded the research in two institutions – a county Probation Department and a group home for incarcerated teen mothers. POLITICS OF PATRIARCHY IN THE EVERYDAY LIFE OF THE STATE In early 1992. To do this. In the first study. It was here. I imagi- ned that the research would centre on the lives and experiences of female clients. In the second study. my inquiries would be grounded in the everyday lives of women in the juvenile justice system. this implied moving beyond the confines of my specific field sites to expose the structural forces that impinged on their institutional rela- tions. the research was to examine female clients’ understanding of the power exerted over them. Most of them denied being part of this system altogether. Almost without exception. the extension beyond the local enabled me to unearth the broader social forces that worked to pit women against one another in the everyday life of the state.

Abramovitz 1988). the main threat to young women was not men but the government. it became clear that these facilities were organized to reflect these messages. was an African American woman in her early fifties. they established house rules geared toward the demonstration of autonomy and they manipu- lated the girls’ relationship to their own babies (who also lived in the house) so as to convince them that one had to be self-sufficient to mother properly. these facilities problematized female dependency and set out to socialize their clients against the currents of private and public patriarchy. this meant the youth prison. The group home staff was equally inventive. As I wrote in an early set of field notes: ‘Could these be feminist institutions? . NEGOTIATING POWER AND EXPERTISE IN THE FIELD 289 were punitive and coercive. According to them. Over and over again. For them. POs presented their work as protective and just. Carol Jackson. For them. Her primary agenda for the more than 100 young women under her supervision was to enhance their independence. this time of ‘public’ patriarchy and its threat to women’s well-being (see Brown 1981.. Zaretsky 1982. They were concerned about how the teen mothers under their supervision seemed reliant on state pro- grammes and assistance. The main PO that I worked with. The staff also believed it was their responsibility to protect young women from the punishment of this other institution. In short. self-confidence and motivation. Carol instructed her clients that a reliance on men would bring them down. In both cases. For Carol. POs continually lectured their clients about the dan- gers of men. The group home staff worked with similar ideas about the importance of female independence and autonomy. female state actors claimed to work in opposition to the punitive practices of the system. Her arguments could have been taken from an introductory textbook dis- cussion of ‘private’ patriarchy and its dangers for women (see McIntosh 1978. Boris and Bardaglio 1983. not only were men unreliable and irresponsi- ble. Their arguments could also have been taken from an introductory textbook discussion.. the California Youth Authority. The group home staff articulated a similar distrust of the system. These arguments about the source of and solution to women’s problems were not simply rhetorical. but also they would squash women’s strength and drive them to crime. They saw it as their job to keep young women out of other parts of the system and to serve as a buffer against its brutality. They constantly lectured the girls about the limitations of government and state assistance. Could I have stumbled on a feminist arm of the state?’ . Perhaps even more surprising were these women’s agendas for their female clientele – agendas that appeared to be quite feminist in orientation. this meant one main thing: to make her female clients less dependent on men. such reliance endangered women: it created a ‘cycle of institutional dependency’ that undercut women’s initiative. they sought to instill self-esteem in their female charges by crea- ting a buffer between them and their boyfriends (or homeboys) and by steering these young women toward their female networks and kin. After spending months observing interactions among these women. Burnstyn 1983).

the teen mothers in the group home protested the staff’s attempts to make them self-sufficient: they mobilized state actors from other facilities to exert power over the staff. These institutional patterns of control and resistance presented me with a series of research dilemmas. passion and feistiness with which they waged it. and they struggled to develop an analysis of what could help these young women. female clients presented a serious challenge to this definition of the situation. Clients’ views about what they needed to survive in everyday life prompted them to resist. Similarly. I was pulled in competing directions. The young women under Carol’s supervision reaffirmed what they believed to be the benefits of their heterosexual relationships: they became more steadfast in their defence of their homeboys. I might have answered such questions in the affirmative. still others viewed men as a source of pleasure in an otherwise painful existence. The teen mothers in the group home were also troubled by the staff’s preoccupation with self-sufficiency. They saw nothing wrong with the use of government assistance: many of them interpreted state support as critical for those with- out the skills to make it on their own. Their resistance also had to be taken seriously. My insistence on listening to women’s voices placed me on both sides of the institutional divide. while others appropriated state sup- port to counter a series of institutional injustices and inequalities. the picture became far more complex. as young women frequently ended up more committed to relationships that may have undercut their well-being in the long run? . In fact. I understood the concerns of POs and the group home staff. they highlighted their femininity to demonstrate what they accrued from it in everyday life. Indeed. they were different from those working in other parts of the system. On the one hand. they accentu- ated exactly those attributes that the women supervising them were so trou- bled by. On the other hand. And when I listened to them. As a feminist researcher. Female clients in both facilities saw nothing emancipatory about the discourses or the practices of the women supervis- ing them. But there was another set of women’s voices struggling to be heard. As a feminist scholar. they even established a welfare club to contest what they considered the staff’s condescending attitude about public assis- tance. they began to behave in precisely those ways that the women in charge of them found so problematic. how was I to make sense of these competing perspectives? And how was I to evaluate the outcomes of these battles. these young women ended up reinforcing precisely what the POs and the group home staff sought to undercut.290 POWER. In doing so. Both sites were comprised of diverse groups of women who were allied and divided in complex ways. PARTICIPATION AND EXPERTISE Had I listened only to the women in charge of these facilities. particularly given the ingenuity. they were deeply committed to the girls they supervised. The young women under Carol’s control found their relationships with men to be unproblematic: some of them looked to men for protection in their violent neighbourhoods. others turned to men as an important source of financial support in their impoverished communities. even their under- standings of these alliances and divisions varied.

their agendas were rooted in their need to survive in everyday life. these women were situated in a bifurcated juvenile justice system. POs looked at these communities and saw men as the biggest threat to their girls’ well-being. Nor could I imagine using reflexivity as an analytical tool by turning inward and socially locating my own knowledge claims. It would have been quite easy for me to deem female state actors as out of touch with the plight of their clients. Hence. This enabled me to understand further these women’s competing definitions of the situation. First. And this frequently pushed them to defend their relationships with men and/or public assistance programmes. For their part. This leads to the second way that I extended the research beyond the con- fines of these institutional interactions. I connected them to the surrounding inner-city community. reflexivity seemed to be of little help: it was just not viable to turn the research into a collective project of understanding or dialogue. And I shudder to think of the politics involved in bringing these sides together to listen to one another. But that would have ignored the larger context in which they worked. Yet the girls’ experiences in these communities . expos- ing them to their clients’ mockery and leaving them vulnerable to attack. In short. instead of turning inward. Their agendas for their clients then reflected this institutional reality. In effect. Indeed. I could not grapple with these research dilemmas by using reflexivity as either a research strategy or an analytical tool. they had relatively limited experience in the system. largely borne out of the deinsitutionalization movement of the 1970s. informing the staff of their survival strategies and thus teaching them how to control the girls more effectively. I grappled with these dilemmas by extending the research outward. They knew quite well that their girls would suffer if channelled into other parts of the system. leading POs to attack private patriarchy and the group home staff to target public patriarchy. NEGOTIATING POWER AND EXPERTISE IN THE FIELD 291 Moving beyond the everyday For a variety of reasons. it would have been nearly impossible to control the social power that my analysis could have wielded. I broadened my analysis to encompass the structural context within which these institutional battles occurred. Or it could have had adverse effects on the young women. clients were not as aware of this broader institutional context. My analysis may have had unintended effects on the POs and group home staff. Instead. As a research strategy. I began to see their contrasting agendas as rooted in very different views of inner-city life. In these deeply divided institutions. In addition to contextualizing these interactions in the structure of the justice system. their institutional positions provided them with particular views of what threatened their clients. a way of retreating from a complex social reality into the subjective realm of autobiography. It also felt like an escape. while the group home staff saw danger in their girls’ reliance on public assistance. my position in the field was so variable that it was nearly impossible to determine how who I am shaped my definition of the situation (Haney 1996). As I have discussed elsewhere. This extension implied two moves.

It permit- ted me to take different women’s voices seriously. I chose two districts with demographic parallels to larger Hungarian trends. by the mid-1990s. In the end. In this way. Thus. I weaved these lessons into my Hungarian study. these women’s institutional interactions were pro- pelled by different positions in and evaluations of the broader urban context. from the onset of my gradu- ate studies. Had I turned inward or remained at the local level. POLITICS OF EXPERTISE IN THE EVERYDAY LIFE OF THE STATE Shortly after I completed my research on the California juvenile justice system. I had planned to locate my dissertation project in Hungary. It was in this area of the world that my interest in gender and the state first emerged: as an exchange student in Hungary in the mid-1980s. PARTICIPATION AND EXPERTISE taught them about the usefulness of heterosexual relationships and of access to state support.292 POWER.and micro-level embodiments of the state and the possible disjunctures in their gender regimes. I also embarked on this study with a far more cautious atti- tude toward ‘reflexivity’ as a research practice and an analytical tool. For all of the reasons discussed above. Nor would I have grasped the complicated ‘lineaments of oppression’. my understanding of the relationship between gender and the state had been complicated. In Dorothy Smith’s terms. it was necessary for me to leave the realm of the everyday in order to problematize its relations of the ruling.’ I learned a similar lesson about the practi- cal limitations of reflexivity in my second ethnography of the state. thus ensuring that my findings were as generalizable . Hence. I embarked on a study of the Hungarian welfare state. setting out to examine the state’s gender regimes as articulated in both its national policies and institutional practices. I would not have captured such complex relations. And it allowed me to understand where each side was coming from and the social factors that shaped their perspectives. As Smith put it in Chapter 1 in this volume: ‘The work of the sociologist is to discover these relations and to map them so that people can begin to see how their own lives and work are hooked into the lives and work of others in relations of which most of us are not aware. My California research had sensitized me to the macro. I first became aware of the important role of the state in the constitution of gender rela- tions. moving beyond the local world of juvenile justice offices enabled me to resist choosing sides in their institutional divides.3 I had been preparing for this latter research for some time. Yet. especially those ema- nating from the structure of the justice system and of inner-city communi- ties. The goal was to analyse the Hungarian welfare state as a layered entity com- prised of multiple apparatuses and gender regimes. Transitional relations of the ruling I began this project by locating two Budapest locales within which to ground the research.

scope and context. this discourse of expertise was highly politicized. In fact. For . their definition of the situation did not involve analyses of the dangers of female dependency. Their institutional battles then centred on who was better able to assess and remedy this deprivation. Within these locales. Thus. and social problems. individual deprivation. they thought they were protecting their clients’ well-being. Others pointed to their education and training as social workers to proclaim that they were best equipped to com- bat client poverty. as the research progressed. In effect. NEGOTIATING POWER AND EXPERTISE IN THE FIELD 293 as possible. this research project had a far broader scope than my California study. state actors differentiated themselves from one another. Thus. At the same time as these institutional dynamics resembled those I encountered in California. I arranged to conduct ethnographic work in three key welfare institutions: child protective services (Gyámhatóságok). Here. in Hungarian welfare institu- tions. I mediated among a total of six institutions in two locales. As in the California study. I was immediately struck by the similarities between the issues plaguing the Hungarian and California state institutions. In this work. child guidance centres (Nevelési Tanácsadók) and family support services (Család Segitö Szolgálatok). I began to see this denial of clients’ gender positions and identifications as a central thread that ran through all of these institutions (Haney 1997). too. As caseworkers adminis- tered means tests and distributed poor relief. Gordon 1994). creating complex alliances and divi- sions between female state actors and their female clientele. Unlike other cases in which